Skip to main content

Full text of "Transactions of the Royal Society of Edinburgh"

See other formats











lo. I. 


January 10, 1908. 

No. XIII. 




November 18, 1907. 





December 16, 1907. 





December 21, 1907. 





December 21, 1907. 





January 24, 1908. 





February 10, 1908. 





February 20, 1908. 





March 27, 1908. 





May 1, 1908. 





April 21, 1908. 





June 16, 1908. 

„ XXIV. 


July 8, 1908. 
July 4, 1908. 
September 2, 1908. 
August 25, 1908. 
October 16, 1908. 
October 15, 1908. 
November 23, 1908. 
November 24, 1908. 
December 24, 1908. 
December 16, 1908. 
January 23, 1909. 
February 18, 1909. 


PART I. (1907-08.) 


I. The Glaciation of East Lothian South of the Garleton Hills. By 
Professor P. F. Kendall and E. B. Bailey, B.A. (With Four 
Plates, Map, and Seven Text-figures), ..... 1 

II. The Plant Remains in the Scottish Peat Mosses. Part III. — The 
Scottish Highlands and the Shetland Islands. By Francis J. Lewis, 
F.L.S., Lecturer in Geographical Botany, University of Liverpool. 
(With Four Plates), . . . . . . .33 

III. On the Source of the Amniotic and Allantoic Fluids in Mammals. By 

D. Noel Paton, M.D., B. P. Watson, M.D., and James Kerr, B.Sc, . 71 

IV. Report on Fossil Fishes collected by the Geological Survey of Scotland 

from Shales exposed on the Shore near Gullane, East Lothian. By 
Ramsay H. Traquair, M.D., LL.D., F.R.S. (With Two Plates), . 103 

V. The Marine Mollusca of the Scottish Natioyial Antarctic Expedition. 
By James Cosmo Melvill, M.A., F.L.S., F.Z.S., and Robert Standen, 
Assistant Keeper, Manchester Museum. (With a Plate), . . 119 

VI. The Pycnogonida of the Scottish National Antarctic Expedition. By 

T. V. Hodgson, F.L.S. (With Three Plates), . . . .159 

VII. Scottish Rotifers, collected by the Lake Survey (Supplement). By James 

Murray. (With Two Plates), . . . . . .189 

PART II. (1908-09.) 

VIII. On the Histology of the Ephedrce, with Special Reference to the Value of 
Histology for Systematic Purposes. By R. J. D. Graham, M.A., B.Sc. 
(With Three Plates), ....... 203 



IX. On the Fossil Osmundacecs. Part II. By R. Kidston, F.R.S. L. & E., 
F.G.S., Foreign Mem. K. Mineral. Gesell. zu St Petersburg ; and D. T. 
Gwynne-Vaughan, M.A., F.L.S., Lecturer in Botany, Birkbeck College, 
London. (Plates I. -IV.), . . . . . .213 

X. The Medusce of the Scottish National Antarctic Expedition. By 
Edward T. Browne, Zoological Research Laboratory, University 
College, London. (With Two Plates), . . . .233 

XL On q-Functions and a certain Difference Operator. By the Rev. 

F. H. Jackson, M.A., ....... 253 

XII. TJie Evolution of the Eyebrow Region of the Forehead, with Special 
Reference to the Excessive Supraorbital Development in the 
Neanderthal Race, By Professor D. J. Cunningham, F.R.S. 
(With Three Plates), ....... 283 

XIII. On the Age of the Old Red Sandstone of Shetland. By John S. 

Flbtt, M.A., D.Sc., ....... 313 

XIV. On Fossil Fish-remains collected by J. S. Flett, M.A., D.Sc, from the 
Old Red Sandstone of Shetland. By Ramsay H. Traquair, M.D., 
LL.D., F.R.S. (With Two Plates), . . . . .321 

XV. On the Lamellibranch and Gasteropod Fauna found in the Millstone 
Grit of Scotland. By Wheelton Hind, M.D., B.S., F.R.C.S., F.G.S. 
(With Two Plates), . . . . . . .331 

XVI. On a New Species of Dineuron and of Botryopteris from Petty cur, 

Fife. By R. Kidston, LL.D., F.R.S., F.R.S.E. (With One Plate), . 361 

XVII. The Craniology, Racial Affinities, and Descent of the Aborigines of 
Tasmania. By Principal Sir William Turner, K.C.B., D.C.L., 
F.R.S. (With Three Plates), . . . . .365 

v/ XVIII. Les Holothuries de V Expedition Antarctique Nationale Ecossaise. 
Par Clement Vaney, Maitre de conferences de Zoologie a la Faculte 
des Sciences de Lyon. (Avec cinq planches), . . . 405 

PART III. (1908-09.) 

XIX. On Lepidophloios Scottii (a neiv species from the Calciferous Sand- 
stone series at Pettycur, Fife). By Wm. T. Gordon, M.A., B.Sc, 
Carnegie Research Scholar in Geology, Edinburgh University. (With 
Three Plates), ....... 443 



XX. An Investigation of the Seiches of Loch Earn by the Scottish Lake 
Survey. Part III. — Observations to Determine the Periods and 
Nodes. Part IV. — Effect of Meteorological Conditions upon the 
Denivellation of Lakes. Part V. — Mathematical Appendix on the 
Effect of Pressure Disturbances upon the Seiches in a Symmetric 
Parabolic Lake. By Professor Chrystal, . . . .455 

XXI. The Structure of Turrilepas Peachi and its Allies. By F. R. Cowper 

Reed, M.A., F.G.S. (With a Plate), . . . . .519 

XXII. Asteries, Ophiures et Echinides de V Expedition Antarctique Nationale 
ficossaise. Par le Dr. R. Koehler, Professor de Zoologie a l'Universite 
de Lyon. (Avec seize planches), ..... 529 

XXIII. On the Fossil Osmundacece. Part III. By R. Kidston, LL.D., 

F.R.S. L. & K, F.G.S., Foreign Mem. K. Mineral. Gesell. zu St 
Petersburg; and D. T. Gwynne-Vaughan, M.A., F.L.S., Lecturer 
in Botany, Birkbeck College, London. (Plates I.-VIIL), . . 651 

XXIV. A Monograph on the general Morphology of the Myxinoid Fishes, 

based on a study of Myxine. Part III. — Further Observations on 

the Skeleton. By Frank J. Cole, B.Sc. Oxon. (With One Plate), . 669 

Index, .......... 683 





^ x »- ■ o e< > 



I. The Glaciation of East Lothian South of the Garleton Hills. By Professor P. F. Kendall and 

Mr E. B. Bailey, B.A. ("With Four Plates, Map, and Seven Text Figures), . . 1 

(Issued separately January 10, 1908.) 

II. The Plant Remains in the Scottish Peat Mosses. Part III. — The Scottish Highlands and the 
Shetland Islands. By Francis J. Lewis, F.L.S., Lecturer in Geographical Botany, 
University of Liverpool. (With Four Plates), . . . . . .33 

(Issued separately November 18, 1907.) 

III. On the Source of the Amniotic and Allantoic Fluids in Mammals. By D. Noel Paton, M.D., 

B. P. Watson, M.D., and James Kerr, B.Sc, ...... 71 

' (Issued separately December 16, 1907.) 

IV. Report on Fossil Fishes collected by the Geological Survey of Scotland from Sltales exposed on 

the Shore near Gullane, Fast Lothian. By Eamsay H. Traquair, M.D., LL.D., F.E.S. 
(With Two Plates), . . . . . . . . .103 

(Issued separately December 21, 1907.) 

V. The Marine Mollusca of the Scottish National Antarctic Expedition. By James Cosmo 
Melvill, M.A., F.L.S., F.Z.S., and Eobert Standen, Assistant Keeper, Manchester 
Museum. (With a Plate), . . . . . . . .119 

(Issued separately December 21, 1907.) 

VI. The Pycnogonida of the Scottish National Antarctic Expedition. By Mr T. V. Hodgson, F.L.S. 

(With Three Plates), ......... 159 

(Issued separately January 24, 1908.) 

VII. Scottish Rotifers, collected by the LaJce Survey (Supplement). By Jamhs Murray. (With 

Two Plates), . . . . . . . . .189 

(Issued separately February 10, 1908.) 



Price Twenty-one Shillings and Tenpence. 


I. — The Glaciation of East Lothian South of the G-arleton Hills. # By Professor 
P. F. Kendall and Mr E. B. Bailey, B. A., of H.M. Geological Survey. Communicated 
by Dr John Horne, F.R.S. (With Four Plates, Map, and Seven Text Figures.) 

(Read May 20, 1907. MS. received August 23, 1907. Issued separately January 10, 1908.) 

The glaciation of East Lothian has already been dealt with separately by Professor 
John Young, who, when attached to the Geological Survey, treated this subject in the 
memoir of the district published in 1866. His account of the phenomena, striae, 
boulder clay, dry valleys, etc., is concise and clear, but he does not enter deeply into a 
discussion of causes. A guarded reference to submersion to account for the presence of 
an erratic boulder of carboniferous sandstone at a height of 1500 feet above sea-level 
shows in fact that he had not arrived at a full conception of the possibilities involved in 
glaciation by land ice. He was, however, quite definite in ascribing many important 
erosion effects in the district to the work of an " ice-stream." In this as in most other 
points he had been forstalled by Sir Archibald GEiKiE,t who had already described the 
immediately adjoining area to the south. 

Since Dr Young's time no one has made a special study of the district. In common 
with the rest of Scotland, our knowledge has of course advanced with increased experience 
won in more extended fields. The writings especially of Sir Archibald and Professor 
James Geikie have fully shown the connection of the local phenomena of the county 
with the wider problems involved in the glaciation of the country considered as a whole. 
With regard to the results produced by the flow of land ice across the district we find 
but little that is new to add. 

It is different, however, in respect to the phenomena of the retreat of the ice sheet. 
These have never been referred to in other than indefinite terms so far as this particular 
region is concerned. The special feature of our communication is therefore the treatment 
of this part of the subject.]: As the historical development of research in this direction 
has, however, been recently discussed in some detail, § we have not thought it necessary 
to re- state it here. 

We must further express our indebtedness to various members of the Geological 

* Published, so far as Mr Bailey's observations are concerned, by permission of the Director of H.M. Geological 

t Geology of East Berwickshire, 1863. 

% J. Geikie has already given illustrations of the main phenomena which accompany the retreat of an ice- 
sheet from a hilly country. Great Ice Age, 1894, ch. xiv. 

§ P. F. Kendall, A.G.J.S., 1902, p. 471, "A System of Glacier Lakes in the Cleveland Hills." 



Survey staff who have mapped or revised the drift of this area. Foremost among these 
is Dr Peach, whose earliest experience of survey work was gained in East Lothian under 
Dr Young. He probably was the first to realise that the Lammermuirs had been 
completely overwhelmed by land ice, and as he worked for several years, at a later date, 
in the area lying south of the watershed, he was able to establish the point beyond any 
doubt. His experience in this. and other matters has always been at our disposal. The 
same may be said of Messrs Clough, Hinxman, Muff, and Crampton, who have been 
engaged in the revision of the area for the Geological Survey, and in close connection 
with whom one of ourselves has been acting. We would further observe that Mr 
Barrow has simultaneously been employed in investigating the allied problems of the 
Garleton Hills, which are situated in East Lothian, but which constitute a naturally 


Fig. 1. 

self-contained area. His description will appear in the new edition of the East Lothian 
memoir. Our best thanks are also due to Mr Lunn of the Geological Survey, whose 
excellent photographs supply the illustrations chosen for this paper. And lastly we 
would like to record our belief that had it not been for the kindly interest and support 
continually given to our work by Dr Hokne, it is doubtful whether the paper would 
ever have been written at all. 

The Maximum Glaciation of East Lothian. 

The glacial history of East Lothian is one of considerable complication. Our 
evidence clearly points to a period of maximum glaciation when the hills were com- 
pletely overridden by a foreign ice sheet ; then came a time when the latter began to 
fail in strength and was unable, even with bulging upper surface, to force a passage for 
itself, but had rather to accept the high plateau region as an independent centre of 


dispersion. Later still the great ice sheet became further enfeebled and shrank away 
back from the hills, merely leaving small local glaciers to linger on for a while in the 
deeper recesses of the upland valleys. 

No doubt similar complications attended the growth of the great ice sheet, so that 
one naturally feels some uncertainty in deciding the proper place in the glacial history 
of the county of any particular event whose record has been preserved. Nevertheless, 
the tale of the maximum stage of glaciation is not hard to decipher in its broad outlines, 
and the approximate direction followed by the bottom currents of the great ice sheet 
will be found indicated on the accompanying sketch map. We would, however, 
recommend the reader to consult Sir Archibald Geikie's * glaciation map of Scotland, 



f o 

r th 

oBass rock 




Fig. 2. 

1. Coal-bearing divisions. 

2. Lower limestone group. 

3. Caleiferous sandstone series. 

4. Carboniferous volcanic rocks. 

5. Carboniferous dolerites. 

6. Old Red Sandstones. 

7. Old Red conglomerates. 

8. Old Red andesites, rhyolites, etc. 

9. Old Red granites, felsites, etc. 

10. Silurian and ordivician. 

11. Transported masses of limestone in 

neighbourhood of Kidlaw and Fala. 

since to understand the progress of events in East Lothian he must have a grasp of the 
main features of the glaciation of the whole central valley. Thus in the west the main 
ice cap of the Southern Uplands proved sufficiently powerful to ward off invasion from 
the Highlands, and the two great ice-streams meeting at the base of the southern hills 
in the plain of Ayrshire turned north-east and south-west to escape either way along the 
central valley. But the Lammermuirs are low, and are bordered to the east by the 
plain of the floor of the North Sea, so that before reaching them the Highland stream 
was able to push its southern companion back upon itself and to creep up obliquely on 

* Scenery of Scotland, 1901, pi. iv. 


to the Silurian Uplands. Once upon the top of the watershed the current turned for a 
while into a direction markedly south of east ; and the same deflection is noticed where 
the congestion in the basin of the North Sea began to make itself felt upon the great 
ice stream issuing from the Firth of Forth ; for the latter had to pass out in both 
directions, and thus the current crossing the eastern Lammermuirs bent round into 
approximate parallelism with the coast towards St Abb's Head. The record of the march 
of the ice sheet is chiefly preserved in the boulder clay, which it deposited as it went. 
Our own observations in this connection relate for the most part to the northern front of 
the Lammermuir Hills and to the district intervening between them and the Firth of 
Forth. They have enabled us in the first place to confirm Prof. Geikie's * record of 
Highland erratics, since the latter occur throughout the whole area and may be 
recognised in good numbers even as far to the south-west as Tynehead. Boulders of 
carboniferous rocks have been carried in company with those from the Highlands far 
out on the Silurian Uplands, and, according to Dr Peach, over the other side into the 
Merse of Berwickshire ; while to the north, along the Old Red and Calciferous Sandstone 
tracts lying at the foot of the hills, the whole country is swathed in boulder clay 
abounding with limestone, coal, and andesite fragments, all indicating a carry from the 
west in a direction oblique to the course of the Lammermuirs. East of the outcrop of 
the Garleton volcanic zone the same tale is told by the appearance of trachytes and 
basalts among the boulders of the drift ; for instance, fragments of the former have been 
found in boulder clay as far to the south-west as White Castle, where the road leads 
across the hills from Garvald. Only two more instances need be cited, and these are of 
special importance, as they serve to indicate in an unusually emphatic manner the 
direction of the ice flow for the districts which they represent. At Port Seton and 
along the coast to the east certain conspicuous boulders of conglomerate occur which 
resemble the Craigmillar conglomerates of the Edinburgh district, and Mr Clough has 
met with a train of similar erratics alone the coast for a mile east of Dunbar. This 
would indicate an average direction of transport of E. 5° N. across the intervening 
district. The second example illustrates the deflection of the course of the ice sheet 
when once it had crossed the watershed. Dr Peach has traced the carry of the granite 
boulder derived from the Priestlaw mass, and found that they indicate an ice flow in a 
south-easterly direction towards Longformacus. 

The evidence just considered may usefully be supplemented by reference to the 
erosion effects produced by the great ice sheet. Strise have been but poorly preserved, 
and all records come from the low-lying districts.t They may be tabulated as follows, 
from west to east ; — 

Kippie Law (half a mile N. of Traprain Law), E. 

Traprain Law (steep south-western face), E. 10° S. 

* Great Ice Age, 1894, p. 195. 

t Striae in the Garleton Hill area are abundantly preserved, and are well known. Their direction is entirely in 
accord with that indicated in the sketch map. 


Dunbar, E. 5° S. (Mr Muff). 

Catcraig Quarry (two miles E. of Dunbar), E. 1 5° S. (Professor Young). 

Shore close to Catcraig, E. and E. 10° N. (Mr Clough). 

A striated pavement, Thorntonloch (5 miles S.E. of Dunbar), E. 10° S. (Professor 

St Abb's Head, E., 25° S. (Professor Young). 
Professor Young* further noticed that the minor inequalities of the low ground have a 
tendency to run in lines almost parallel with one another, a fact which is well illustrated 
in a shaded copy of an Ordinance one-inch map (Sheet 33). He attributed this to the 
result of direct sub-glacial erosion, and noticed that if a line be taken joining Lothian 
Edge to Tynemouth (see general map), it separates the lowlands into two districts. 
West of the line the grooving runs E.N.E. or E. by N., but on its seaward side the 
features " become much more easterly and even a few degrees south of east." This type 
of ice modelling is not so widely developed in East Lothian as in many parts of 
Scotland,t but it is well shown in the district south and east of Haddington. 

The direction of ice flow indicated by the modelling agrees approximately with that 
of the few strias observed, and thus the erosion effects as a whole, where recognisable, 
serve as a useful check upon the evidence afforded by the deposits. They clearly 
demonstrate that in the low-lying parts of the county the ice sheet followed roughly the 
trend of the central valley, until, debouching upon the North Sea plain, it was forced 
to take a somewhat southerly course, skirting the termination of the Southern Uplands. 
The district lying south and east of Haddington affords a good illustration of the 
special characteristics of an ice-dressed surface upon a large scale. The topography 
induced is essentially that of a drumlin country. Long low mounds or ridges run 
parallel with one another, merely separated by shallow broad-bottomed grooves or valleys. 
If any particular ridge be followed it will presently be observed to sink slowly and 
gradually until the grooves on either side meet about its termination ; others have, 
however, by this time arisen situated en echelon to their failing companion, and so the 
group persists though its individual members prove inconstant. The distinction between 
these mounds and true drumlins depends upon the fact that they have been fashioned 
in solid rock and are merely coated over with boulder clay, very thin along the crests 
of the ridges, instead of being entirely formed of the latter. Wherever an isolated resis- 
tant mass occurs, this type of sculpture naturally becomes accentuated, and Traprain 
Law furnishes a familiar example of " crag and tail." By taking note of such where they 
occur, it becomes evident that the whole grooving must be referred to an ice sheet 
moving towards the E.N.E. In following this course the current flowed from the area 
occupied by sedimentary rocks south of Haddington over the igneous rocks occurring 
south of Traprain Law. The phenomena marking this passage from the soft rocks to 

* The Geology of East Lothian, 1866, pp. 63, 64. 

t Cf. Sir A. Geikie, Geology of East Berwickshire, 1863, p. 52, and J. Goodchild, Glacialisti Magazine, vol. iv. 
pt. 1, 1896, p. 1. 


the hard are very beautiful. The grooves persist in their original direction, but grad- 
ually rise to surmount the obstruction, and upon the other side sink as gradually to 
their former level. A small patch of alluvium near Morham Church serves to mark the 
original presence of a shallow tarn excavated along one of these grooves just east of 
where the change of slope commences. 

Certain of the grooves south of Haddington have been employed as temporary 
watercourses at a later period, but the work of the streams which thus made use of 
them was chiefly the deposition of alluvium upon the broad bottoms of the hollows. 

Elsewhere the sculpturing is not so well developed as in the district just described, 
but round the termination of the hills it serves to indicate the deflection of the ice 
current which has already been referred to. 

Let us now return from the consideration of the direct effects of sub-glacial erosion 
to the correlative phenomenon of deposition. The district furnishes three very fine 
examples of large transported masses of limestone. The most conspicuous is that of 
Kidlaw, a mass one-third of a mile long and a quarter broad and well exposed in 
extensive disused quarries. The view that this great mass might really be a boulder 
was first suggested some years ago in conversation by Mr Howell, and the evidence 
for this interpretation is threefold. 

1. The few exposures in the immediate neighbourhood and general considerations 
as to the geological structure of the district, point to the conclusion that the limestone 
is resting on sediments of Upper Old Red or early Calciferous Sandstone age, with 
associated igneous rocks. The limestone, moreover, has been recognised by 
Dr Crampton as belonging to the middle and not the bottom of the Lower Limestone 
group of the Scotch Carboniferous sequence. Under these circumstances one has to 
account for the apparent absence of a great thickness of strata, all included in a 
conformable sequence and well developed in the near neighbourhood. 

2. The whole mass is in a completely shattered state. Mr Howell * in his 
original description attributed this feature to the proximity of the Lammermuirs fault. 
But the longest axis of the " outlier" is directed away from the fault, and it would be 
very difficult, therefore, to maintain this interpretation ; moreover, since Mr Howell 
mapped the ground, a large basalt quarry has been opened in the immediate vicinity 
and equally close to the fault. The basalt sill is there entirely unshattered, though 
the period of its intrusion certainly greatly antedated the faulting. 

3. There is a great abundance of limestone blocks, both large and small, in the 
boulder clay of the adjoining areas, whether it rests upon Silurian, Old Red, or 
Carboniferous strata. 

The second example is recorded by Dr Crampton at Marl Law quarry, a mile N.W. 
of Fala. It consists of a contorted mass of limestone and marl, lying not very far from 
the natural outcrop of the group. It may perhaps not be without significance that 
the Kidlaw and Marl Law masses lie along the same line of glacial transport. 

* Geology of East Lothian, 1866, p. 55. 


The third example occurs at Woodcotte Park, a mile to the east of Fala. The 
limestone forms two little knolls known as the Little Hill and Meikle Hill respectively. 
Here, as at Kidlaw, the field evidence strongly suggests that the limestone is resting 
on the Upper Old Red Sandstone, only in this case upon the south side of the 
Lammermuir fault. The exposure was known some sixty years ago to Charles 
Maclaren * and David Milne HuME,t who regarded it as belonging to a natural outcrop. 
Maclaren remarks, however, that he saw none in place. The evidence for believing 
that these knolls consist of transported limestone is therefore identical with that put 
forward in respect to the Kidlaw example, and the probability of the interpretation is 
heightened in that it is not invoked merely to explain a single occurrence. It is 
noteworthy that Mr Howell in his mapping long ago did not treat these smaller 
masses as being in place, although one of them was formerly quarried and burnt 
for lime. 

The carboniferous limestone seems in certain other places to figure more largely in 
the drift than might at first be expected ; this suggests that in pre-glacial times the 
more important beds of limestone may have given rise to conspicuous little escarpments 
with long, exposed dip slopes, and that these furnished an abundant supply of blocks 
during their degradation by the ice sheet, while at the same time they may have 
afforded favourable opportunity for the transport of big masses. 

The Retreat of the Ice Sheet. 

We have now examined some of the results of the action of the ice sheet during its 
maximum development, but perhaps even more interesting still are the circumstances 
of its retreat. When gradually it was forced to give place to other agents of erosion, 
it yet was able for a while to impose special conditions upon their operation. Thus 
step by step the enveloping ice shrank back to leave the Lammermuirs standing like a 
stone in the midst of melting snows ; and stage by stage the memorials of this retreat 
were furnished by the obstructed drainage of the hill country, joined by the waters 
issuing from the glacier itself : There are whole suites of channels to bear witness to 
the former presence of swollen torrents, linking together the long chains of lakes, which 
then rested in temporary support between the steep ice margin and the slopes of the 
hill country to the south ; and there are also the deposits which accumulated in these 
lakes and at many an intervening point where the connecting streams had not as yet 
cut gorges for themselves. 

One can even recognise by a careful investigation of the evidence that there were 
times when the ice sheet renewed its vigour. One can trace its oscillatory re-advances : 
here it spreads its boulder clay over the floor of a lake where previously sand and silt 

* Sketch of the Geology of Hfe and the Lothians, 1839, p. 74. 
t Coalfields of the Lothians, lb39, p. 103. 


had been collecting, and these re-advancing across the path of a glacial stream it drives 
the occupant back again into some deserted channel further up the side of the hills. 

A point of special interest is the fact that all these effects are restricted below a 
more or less clearly defined upper limit (see fig. 1). Marginal streams and temporary 
lakes have scarcely left a trace of their former presence in the higher portions of the 
hills. Yet it can be demonstrated that this does not mark the upper limit reached by 
an invading ice sheet, but rather the level at which the retreating ice parted company 
with its lineal descendants, the little glaciers of its high-level valleys. 

Moreover, the district has furnished fresh evidence regarding the relations which 
locally prevailed between land, ice, and water, when the ice sheet had already withdrawn 
into the low ground ; and, probably what will attract more general interest, it has 
supplied a number of very beautiful instances of a special type of stream capture, which 
has greatly influenced the plan of the glacial drainage system. Without entering into 
detail at present, it is sufficient to say in this connection that there are many cases 
where a tributary stream has thrown out a cone or delta into a deserted glacial drainage 
channel and has thus succeeded in establishing a watershed within the latter.* 

In order to illustrate the various aspects of the subject alluded to above, we will 
arrange the evidence under the following heads : — 

I. Phenomena illustrating the general conditions which attended the retreat of the 

great ice sheet. 
II. Phenomena illustrating the oscillatory nature of the retreat. 
III. The upper limit of these marginal phenomena set by the confluence of the 

local glaciers with the great ice sheet. 
IV. The influence of corroms t and bridge-deltas in modifying the glacial drainage 

V. Evidence regarding the level of the sea at the time of the retreat of the ice 

I. Illustrations of the Phenomena Attending the Retreat 
of the Great Ice Sheet. 

East Lothian illustrates extremely well two distinct though cognate aspects of the 
phenomena which attended the retreat of the great ice sheet. One notices everywhere 
records of the operation of water under special conditions of restraint ; in one place the 
manifestation may take the form of some monumental piece of erosion, in another of 
deposit, but everywhere the phenomena point to the action of water, and of water 
impounded in front of the retreating ice sheet. One of the consequences of this 

* Dr Crampton independently arrived at precisely this interpretation in accounting for the partial reversal of 
the drainage in the Borthwick dry valley, to which further reference will be made in the sequel. 

+ Corroin (cothrom) is a Gaelic word used in place names in the Ardgour district of Argyllshire, to denote a. 
delta-watershed. Its literal meaning is a "balance," and it is intended to illustrate that a stream issuing upon such 
a cone has the chance of flowing either the one way or the other. 


condition of affairs that has continued to this day is the anomalous drainage system 
possessed by East Lothian and the adjoining coastal district of Berwickshire, the 
existence of which has long been recognised. In fact, no better introduction to our 
whole subject can be desired than the descriptions given, now more than forty years; 
ago, by Sir Archibald Geikie* and Professor J. Young. t Besides giving a general 
account of the anomalous drainage system, both these authors have singled out for 
special notice examples which seem abnormal even in the strange company of their 
fellows, and which, as we shall attempt to show later, owe their special peculiarities to 
the processes of corrom formation. Sir Archibald's account deals with the Berwickshire 
area, and runs as follows : — 

" Besides the ravines .... which are plainly the work of rivulets, there are certain 
other narrow glens, defiles, or channels (for they vary not a little in size) in which there 
are either no streams, or runnels so tiny and local that they cannot be imagined capable 
by their own unassisted operations of hollowing out such marked excavations as those in 
which they now flow. On the south side of St Helen's Church, for example, there is a 
deep winding ravine, quite dry, running along the side of the declivity which here 
descends gently to the edge of the sea cliff. It is open at both its eastern and western 
extremity, thus running across the neck of a promontory. With the present configura- 
tion of the land no stream could ever have flowed along this ravine. Other, but less 
marked, examples may be seen along the sea front eastwards towards St Abb's Head. 
Along with these features ought probably to be classed those valleys which go right 
across the watershed in different parts of this district. Keference has already been made 
to the valley traversed by the North British Railway J as affording a good illustration." 
He then proceeds to give a description of the remarkable course followed by Edmonds 
Dean Burn, which enters the railway pass " within a few yards of the flat watershed," the 
latter being one of the corroms to which special attention will be directed later ; and he 
concludes his account with the suggestion that some of the peculiarities of this drainage 
system may be due to the modifying influence of direct glacial erosion. 

Professor Younu, however, in his description of the East Lothian examples, carefully 
refrains from any expression of opinion as to the mode of origin of the valleys. 
Having described the glacial grooving, he proceeds : — " In connection with these glacial 
hollows fall to be considered certain valleys near the base of the hills, which 
seem to belong to a system different from that of the present day. The flat green 
valley § between Deuchrie Dod and Rammer Woods is bounded by steep declivities, 
covered with low brushwood." After describing this abandoned waterway, the finest 
in the district, he passes on to the Presmennan Loch channel, a little to the north, and 
then draws special attention to the peculiarities of the Thurston and Aikengall drainage 
systems, peculiarities which we now regard as a sequel to corrom formation. Returning 

* The Geology of East Berwickshire, 1863, pp. 51, 52. 
t The Geology of East Lothian, 1866, pp. 63, 61. 

X Between Cockburnspath and Grant's House stations, eist of Dunbar. 
§ PL I. %. I. 


to the general aspects of the case, he continues : " In all these cases the course of the 
dry valleys is the same, north-eastwards, and is, therefore, at right angles to the present 
stream courses. It is difficult to explain their formation by means of the small brooks 
which flow in their neighbourhood, and their form is not that which would result from 
glacial erosion. There are many such valleys, some of which are mentioned in the 
Berwickshire memoir ; the same explanation does not, however, remove the difficulties 
in each case ; their solution must, therefore, wait the results of more extended 

AVhen we add that many of the most typical examples are excavated entirely in 
drift deposits, it becomes evident that the anomalies referred to cannot be regarded as 
vestigial inheritances from a pre-glacial drainage system, nor can they be accounted for 
as the consequence of a selective erosion acting along lines of geological weakness. 
From what has already been stated the reader will have gathered that we intend to 
apply the same explanation here as was developed by one of us * in order to account 
for similar peculiarities in the Cleveland district of Yorkshire. 

Sir Archibald Geikie's remark with reference to the channel on the south side of 
St Helen's Church might be repeated a score of times. " With the present configura- 
tion of the land no stream could ever have flowed along this ravine." In the particular 
case he cites, it might perhaps be suggested that the valley has been cut back to by 
recent marine erosion, but such an explanation carries no weight in view of the 
numerous inland examples. The inadequacy of " the present configuration of the 
land " is nowhere better illustrated than in certain cases where stream gorges have been 
cut into the uniform slope of a hill side, following instead of crossing the contour lines. 
The Pressmennan Loch channel furnishes an imposing example. During the initiation 
of this gorge it is necessary to suppose that the ice front itself supplied the northern 
containing wall of. the stream, until the latter had sunk its channel into the solid rock. 

Almost equally impressive is the entrance to the Chesters Quarry ravine, a mile north 
of Garvald. It seems at first sight indeed to have been fashioned in sheer contravention 
to the ordinary laws of Nature, for it originates on an open plane, and straightway leads 
across the imposing spur of Whitelaw Hill. But once in imagination replace the ice 
sheet to the north and the difficulty vanishes. A lake forms behind the rocky ridge, 
and its escaping waters are forced to take the route which at the present time seems so 

These are merely examples chosen well-nigh at random from a host. The Chesters 
Quarry defile may also be noticed here as a representative " dry valley " ; for not only 
was it cut entirely by glacially diverted waters, but being so cut it has in later times 
failed to retain any stream of its own. Naturally, however, a great number of glacial 
drainage channels are still employed as watercourses ; there is in such cases a very 
marked tendency leading to the formation of corroms and the consequent establish- 
ment of short cuts, so that the inadequacy of "the present configuration of the land" 

* P. F. Kendall, "Glacier-lakes in the Cleveland Hills," Q.J.G.S., 1902. 


to account for the origin of the anomalous drainage system becomes all the more obvious 
in that the latter is so conspicuously unstable. 

We shall now pass on to the discussion of the modifications other than corrom 
formation which may be attributed to the streams which at present occupy glacial 
drainage channels ; and as a preliminary step we may allude to a peculiarity of the 
original glacial channels which has proved itself of great importance in this connection. 
They have as a rule been cut to a very gentle gradient, and an example occurs, between 
Kidlaw and GrifFord, which one might more fitly describe as the dry valley of a sinuous canal 
rather than that of a stream. The feature no doubt must be accounted for by the fact 
that at certain intervals these channels were called upon to conduct away large masses 
of water which were able to maintain a current even after they had degraded their beds 
to the local base level. Already several of these valleys have been dammed to form 
reservoirs or ornamental lakes, and the number is always on the increase ; a low dam 
serves to impound a considerable body of water owing to the extremely gradual fall in 
the valley bottom ; and in fact at Pressmennan* it has been found necessary to place a 
dam at the head as well as at the foot of the lake. Small natural tarns and marshes 
are also frequent along the course of the streamless valleys impounded behind screes 
* or deltaic cones. 

The glacial drainage channels still occupied by streams can, as a rule, be readily 
recognised even upon the map, since they form an integral part of the anomalous 
drainage system of the county. The gorge in which Spott Burn lies, and which proved 
so terrible an obstacle to the Scots in their attack upon Cromwell, furnishes an excellent 
illustration within easy walking distance of Dunbar. The burn itself is of very con- 
siderable volume, and as it winds along the broad bottom of the valley it may be seen 
in many places cutting into the steep bluffs on either side. Here at first sight one 
might reasonably expect to find evidence of well-marked deepening of the channel since 
glacial times. But the actual phenomena of the case emphatically negative this 
suggestion, while they enable us to recognise the modification for which the present 
stream has been responsible. A little to the west of Spott the valley, as we follow it 
up-stream, bifurcates,t or rather it is joined from the south by another valley of identical 
type. The remarkable feature is that the direct continuation westwards of the Spott 
valley is entirely streamless,^ for the whole volume of water is supplied by the valley 
entering from the south. Examination of the dry valley proves that it owes its stream - 
less condition entirely to natural causes, and it may confidently be regarded as an 
overflow channel which has been deserted ever since the final retreat of the ice front 
from this neighbourhood. It therefore furnishes a datum from which to reckon the 
greatest possible vertical erosion accomplished by the Spott Burn in post-glacial times. 
The dry valley enters flush with the bottom of the stream-containing valley, or at most 

* The same is being performed for Lammer Loch now that the latter is being extended to form a reservoir. 
+ PI. II. fig. 1. 
I PI. I. fig. 2. 


the difference between the two does not amount to more than four feet, showing clearly 
that the post-glacial vertical erosion has, in this case, been quite trivial. Standing below 
the junction of the two confluent valleys we obtain a good view up both, and are able 
to appreciate at once their similarities and their contrasts. They are both gorges 
bounded by steep walls ; but while the deserted channel has lost some of its original 
abruptness of outline, owing to the collection of rain wash, it evidently retains a feature 
which in the case of its neighbour has been more or less completely obscured by the 
action of the stream. For at every convexity in the course of the dry valley the outer 
■wall is more steeply cut than the inner, showing that the valley was excavated by a 
stream accurately proportioned to the channel it was cutting for itself,* and of quite a 
different order of magnitude to the Spott Burn. 

The role played by the Spott Burn during post-glacial times may now be readily 
appreciated. Since ever its volume has become so greatly reduced it has wound in 
small meanders along the valley bottom, cutting now into the one side, now into the 
other. In this manner it continually supplies itself with as much material as it can 
transport, and thus as the meanders work gradually down stream, the valley is widened 
but nowhere materially deepened. 

Two or three equally striking examples occur in the neighbourhood of Garvald, but 
it is important now to point out that some of the East Lothian streams left in occupancy 
of glacial drainage channels have proved themselves capable of accomplishing a notable 
amount of vertical erosion, and to recognise that in such cases the resultant changes can 
readily be distinguished. The Braidwood Burn,t a little south of Innerwick, is an 
excellent case in point. It is a powerful stream, and has cut for itself a pronounced 
gorge, but the result is far more accurately described as a dissection of the floor of the 
old valley, rather than a deepening of the same. For the work of the recent stream, 
with its greatly diminished volume, has been to cut a gorge that winds to and fro along 
the old bottom, the latter being in large measure still preserved intact. 

It is worth while before leaving the subject of these channels to take note of the 
enormous volume of water which they must at times have been called upon to conduct 
away. They served to carry off the drainage, not only of the Lammermuirs, but also of 
the Moorfoots, and in fact of the whole of the Mid Lothian basin, except in so far as 
leakage was permitted here and there through the fissured margin of the ice sheet. 
During the earlier stages of retreat this western drainage sometimes found a short cut 
across the watershed by the pass between the Moorfoots and Lammermuirs, which has 
been taken advantage of in our own days for the railway leading south to Galashiels. 
But the summit level here is 884 feet, so that for a long period this short cut was not 
available, and the Mid Lothian drainage had perforce to pass right across East Lothian 
and part of Berwickshire to reach the sea. Several routes were successively employed 

* This is a genera] feature of dry valleys. See Glacier Lakes in the Cleveland Hills, p. 483. The type example 
for East Lothian is furnished hy the Danskine Loch channel, two miles east of Gifford. 
t PI. II. fh.'. 2. 


by the waters passing from the one natural drainage area to the other, and one of them 
is of such importance as to deserve special notice. It is the Borthwick dry valley, 
connecting the course of the Gore and Tyne waters, and crossed by the railway just 
south of Tynehead station. The changes which have led to its abandonment will be 
dealt with in a later section, so that we shall only notice here the advantages which it 
owed long ago to its particular situation. The Mid Lothian and East Lothian basins are 
separated by a line of undulating hills, constituting the Eoman camp ridge which 
extends northwards approximately at right angles to the general trend of the margin of 
the retreating ice sheet. The col at the southern extremity of the ridge is not well 
marked, but once the Borthwick channel had been established at this ill-defined pass 
and had been excavated to its present depth, the ice sheet had to retreat fully 2f miles 
further to the north before disclosing any opening of lower altitude across the ridge. 
The Borthwick channel was therefore employed for a protracted period during the 
retreat, and at one time must have served to drain an important lake occupying a 
portion of the Mid Lothian basin. 

We may now pass on for a while to consider briefly the deposits which were, 
produced by the agency of the same constrained waters as cut out the drainage channels 
just described. 

Great masses of sand and gravel encompass the Lammermuir Hills round about, and 
have long been familiar to Scots geologists. Like other accumulations of their kind, 
they appeared to the earlier workers as the memorials of great submergences. 
Professor Young, # however, in his record of the distribution of the deposits, refrains 
from offering any suggestion as to their mode of origin ; and later, Professor GEiKiE,t in 
several brief references, rightly regards them as of fiuvio-glacial origin. 

The glacial age of the gravels is beyond question. They overlie and here and there 
intercalate with the boulder clay. They everywhere contain erratic pebbles, most 
conspicuous among which are fragments of coal assorted into special layers in accordance 
with their low specific gravity, and finally their distribution makes it necessary to 
believe that they formed in proximity to the retreating front of the ice sheet. 

Various factors have been concerned in their production, and we may mention the 
following main sources from which they have derived their material. 

1. The hill country, owing to the melting of its local glaciers, the re-opening of its 
normal lines of drainage, and the rapid waste of the boulder clay covering its steeper 
slopes, supplied a very large proportion of the whole. Along the northern front of the 
hills the contrast between the composition of the gravels and that of the boulder clay 
is rendered very marked by this circumstance. The boulder clay resulted, as we have 
seen, during a passage of the ice sheet from west to east, often from low ground to high, 
while the gravels of this series travelled in large measure from south to north, and 
always from high ground to low. Thus while the boulder clay, carried on to the 
Silurian Uplands, is charged with the debris of the rocks of the central valley, the 

* Geology of East Lothian, 1866, p. G5. t Great Ice Age, 1894, p. 211. 


gravels, on the other hand, which extend on to the latter from the hills, are everywhere 
characterised by pebbles of Silurian grauwacke.* 

2. The excavation of the glacial drainage channels is responsible for another great 
part of the material, especially, perhaps, that which litters the coastal plain. 

3. The ice sheet itself also no doubt directly supplied some proportion of the deposit. 
As to the conditions regulating the accumulation of these sands and gravels, it 

seems certain that for the most part they formed either as the deposits of temporary 
ice-dammed lakes or as gravel spreads extending between the ice and the hills and 
reaching, along the coastal region, far out over the shelving bottom of the sea. 

It is difficult now to distinguish between these somewhat differently accumulated 
deposits, and, in fact, one can only speak with confidence in special cases. It has been 
found impossible to draw a line separating the gravel spread of the coastal plain into 
two portions, according as it formed above or below the ancient high- water mark, or in 
the inland sections to differentiate the material which accumulated in the temporary 
lakes of the period. 

Not only were the original boundaries in the majority of such cases very blurred 
and ill-defined in their nature, but also the deposits have since frequently taken on 
outlines corresponding to a mature stage in erosion. We may illustrate this fact by 
an example from the neighbourhood of Upper Keith ,t north-west of Fala. Several 
interesting sections occur here, which render this the type area for the study of the 
deposits, but their full description will be delayed till the section dealing with the 
oscillatory nature of the retreat of the ice sheet. For the present it suffices that the 
sections show a great mass of evenly bedded sands, with here and there a layer of 
gravel or another of laminated clay or warp. The whole appears as the deposit laid 
down in the tranquil waters of a great lake, but now so deeply eroded that it has taken 
on the moundy form which one naturally associates with a sand district. Here and 
there, indeed, flat stretches do exist which at first sight might be taken to mark an old 
level of the lake waters ; but, as will appear later, these merely indicate the outcrop of a 
more resistant bed of boulder clay which lies intercalated in the sands, and have there- 
fore an entirely different significance. 

With this example in view it would be rash indeed to regard any of the " kame " 
outlines, which are locally developed, as due to the preservation of original forms of 
deposit. In fact, the most conspicuous example of kamiform ridges of gravel in this 
district are afforded near Woodhall, south of Spott, where a section shows that the even 
and almost horizontal bedding of the deposit bears no relation whatever to the external 
form of the mounds, but is sharply truncated by the steep hill face. The interpretation 
is obvious in this particular instance,! for the ridges are part of a gravel-spread which 
formed between the hills and the ice sheet at a time when the glacial waters were being 

* Numerous exposures in the neighbourhood of Upper Keith, for instance, illustrate this point. 

t PI. IV. figs. 1 and 2. 

I We do not suggest for an instant that true kames of deposit do not exist in other districts. 


supplied with such a superabundance of material from west and south that even they 
were unable to cope with it. Later, when this supply had become in part exhausted, 
and the passes to the east had at the same time been deepened, the gravel-spread was 
attacked and kamiform ridges produced. Tn fact, the latter are here nothing more than 
the remnants left between successive glacial drainage channels. The disposition of the 
channels in this case is such as to indicate that they were formed when a re-advance of 
the ice sheet had carried it forward again over the gravels ; but there are equally 
numerous cases where a wide stretch of sand and gravel has been dissected by glacial 
drainage channels, for which such an interpretation is quite unnecessary. In imagina- 
tion place an ice sheet abutting against the front of any hilly tract of country, and 
suppose that it remain stationary for a long time in the same position. The governing 
principle operative throughout will be the tendency of the obstructed drainage to 
furnish itself with a suitably graded channel leading to the nearest available outlet. 
Hollows and gentle slopes will be loaded with sands and gravels, while ridges, which 
rise above the general level, will be breached by cross-cut channels : but, supposing the 
process perfectly continuous, reach after reach, where at first sediments had accumulated, 
would attain, as it were, its saturation point and, refusing further deposit, would hence- 
forth enter upon a career of erosion. 

Under more natural circumstances further complications would be certain to arise : 
thus we have already hinted several times at the oscillations which marked the with- 
drawal of the ice sheet from the Lammermuirs ; it is possible, however, in a first broad 
view, to neglect the results attributable to these oscillations, since they affect matters 
of detail and not of general principle. But even then the account we have sketched 
above is inadequate, since the opportunities for the destruction of sand and gravel 
deposits formed in front of a stationary ice sheet are evidently far less numerous than 
if the ice sheet were slowly and continuously retreating ; thus many a time a once- 
continuous terrace must have become reduced to a mere collection of isolated remnants. 

We are now in a position to read the evidence afforded by the East Lothian sands 
and gravels. A remarkable feature in their distribution is that they are for the most 
part concentrated into two definite and quite distinct areas in which, as we shall 
endeavour to show, they have accumulated under somewhat different conditions. The 
first of these areas is a well-marked belt or zone of maximum accumulation extending 
from Tynehead in the west right round the hills to Oldhamstocks in the east, while the 
second includes what may be termed the coastal spread. 

Examination shows that the first of these two zones or belts is roughly contained 
within limits set by the 700- and 500-foot contour lines respectively ; further, a study 
of the mode of occurrence of these deposits in the field shows that many of the gaps 
which break the continuity of the belt are certainly the result of erosion. In fact, it 
seems tolerably clear that at one time a sand and gravel terrace, similar to that described 
by Prof. Geikie * for the Strathavon Hills, near Glasgow, reached round about the 

* hoc. cit., p. 175, pi. iii. 


northern and eastern Hanks of the Lammermuirs. passing by Tynehead in the west and 
Oldhamstocks in the east. Between these two villages the continuity of the terrace 
was broken by two important spurs, namely, Deuchrie Dod and Cocklaw Hill. Behind 
each of these at one time or another important lakes were doubtless imprisoned, and 
it seems probable, from the nature of the deposits in the Upper Keith district, that 
Deuchrie Dod served temporarily as the eastern limit of a narrow marginal lake fully 
fourteen miles in length. But the history of this terrace was long and varied ; we 
shall presently have to deal with evidences to show that the ice in front of which it 
formed was frequently oscillating to and fro. In fact, it marks much the most important 
halt, if such a name may be applied to a limited though restless wandering to and fro, 
which occurred during the retreat of the ice sheet from East Lothian. Thus it came 
about that the marginal drainage system of this zone reached a mature development 
along the lines indicated above. At first almost the whole extent of the marginal area 
was loaded with sediment, though, of course, the supply varied greatly in amount and 
from place to place ; erosion occurred only across the two spurs already mentioned until 
the general lowering of the base level caused the emergence of minor ridges and the other 
consequences which we have indicated above. But before the ice sheet had finally 
retired, the terrace so formed had been dissected from end to end by marginal streams 
and the work of its demolition had already far advanced. 

Above and below the limits of this great terrace we may recognise others, but never 
on the same grand scale. A well-marked higher terrace may be referred to leading 
eastwards from White Castle towards Deuchrie Dod. Its level is approximately 800 
feet above the sea. 

We shall now consider the other great zone of deposit, that of the coastal spread. 
At Oldhamstocks it comes into contact with the great Lammermuir terrace, but it is 
obvious in the field that the deposits of this locality have an extremely complex history, 
and that those occurring at hio;h levels were formed at a time when the ice sheet must 
have blocked all direct egress to the sea. The whole coastal spread is, in fact, built 
up on this plan, and it owes its continuity to the juxtaposition and overlap of deposits 
formed successively and not contemporaneously. It cannot therefore be regarded as 
an entity, even in the same extended sense in which that term might be used to describe 
the complex Lammermuir terrace. It is merely the gravel spread which formed where 
the mouths of the glacial drainage channels debouched upon the coastal plain, and as 
these came into being in succession from south-east to north-west, following the retreat 
of the ice sheet, it is obvious that the coastal gravels near Oldhamstocks, for instance, 
were formed at a considerably earlier date than those about the mouth of the Tyne. 


II. Phenomena Illustrating the Oscillatory Nature of the Retreat. 

We have already referred to the great accumulations of fine sands, silts, and gravels 
which occur in the Upper Keith neighbourhood. They have all the appearance of 
having been laid down in a great lake, and it seems necessary to suppose that during 
the time of their formation the ice sheet of the lowlands lay in close proximity to 
the north and thus served to dam back the waters in which they accumulated. 
A closer inspection of the many good sections * occurring in this district confirms this 
interpretation, since such an examination affords independent and clear evidence 
of the continued proximity of the ice sheet during the formation of the sands. 

Two of the sections, both of them occurring along the course of the Keith water, 
are specially clear ; the first of them constitutes the Red Scar,+ a little north of Costerton 
cottage, and the second j lies just to the west of the road from Upper Keith to Ormiston. 
Their details are as follow : 

Feet. Feet. 

4. Sand . . . 53 Not exposed in this section. 

3. Boulder clay . . 13 Boulder clay (top not seen) . 6 

2. Sand ... 15 Sand 65 

1. Boulder clay (base not Boulder clay (base not seen) 33 

seen) . . . 37 

Even in the Red Scar the section is incomplete, and the upper sands must be fully 
150 feet in thickness. The thin gravel beds sparingly interbedded with the sands carry 
a conspicuous proportion of grauwacke pebbles derived from the Silurian Hills two 
miles to the south, but the larger blocks of the intercalated boulder clay consist of 
sandstone, etc., so that it seems necessary to suppose that the latter was deposited by 
the lowland ice sheet. In fact, we take its presence to indicate that this glacier was 
not melting away as an inert mass of ice, but that it was ready, when climatic conditions 
favoured, to re-advance on to the floor of the temporary lake which spread out before 
it, and there deposit a covering of boulder clay upon the sands and silts which had 
previously been collecting. The intercalated boulder clay can be traced w r ith almost 
complete certainty for a distance of a mile in a direction parallel with the ice margin, 
and a quarter of a mile at right angles to the same, without any indication of coming 
to an original limit, so that the oscillation recorded here seems to have been of con- 
siderable magnitude. The positions of the best sections in this interesting district are 
indicated by asterisks placed on the general map. 

Many other exposures occur along the Lammermuirs which tell the same tale as 
those of Upper Keith, but not in so impressive a manner ; we need therefore only 
allude here to the readily accessible exposures in the railway cutting soutl of Tynehead 

* The interstratification of sands and boulder clays here was first noticed by Mr Anderson. 
tPl. IV. fig. 2. 
J PI. IV. fig. 1. 



station, before passing on to indicate other lines of evidence which point equally clearly 
to the oscillations in which the great ice sheet effected its retreat. 

In the first place, still dealing, as at Upper Keith, with the deposits of the great 
Lammermuir terrace, we may draw attention to the highly significant erosion forms 
which they locally present. As already noticed at Woodhall, south of Spott, the 
terrace has been so completely dissected by a series of glacially directed stream courses 
that its sands and gravels have now assumed a kamiform aspect, and it is obvious in 
the field that the ice sheet which directed this erosion must have re-advanced after a 
retreat, until it actually came to stand upon the top of the spread of deposits which 
previously had accumulated before it. 

But this is not the only way, or indeed the most important way, in which oscillations 
of a retreating ice sheet may be recorded in the erosion effects which it determines, and 
to make our line of argument more readily appreciated, we have inserted the following 
explanatory diagram. 


600 A B 


500 __ 

Fig. 3. 

The contour map and section in fig. 3 show an ideal case of a spur reaching out 
from a line of hilly country at right angles to the general trend of an ice margin 
retreating towards the north. The ridge has been breached by two overflow channels, 


the one at A, the other at B. It will also be observed that the top of the A channel lies 
at a higher level than the bottom of the B channel, so that the latter cannot have been 
excavated to its present depth when the former began to be cut. This indicates that 
the two were not employed merely during a continuous retreat of the ice front, for 
under such circumstances no channel could have begun to cut north of B until the ice 
sheet had withdrawn beyond the 525-foot contour. It is evident, then, that these 
two were cut, in part at least, during an advance of the ice sheet, so that if there is 
reason to believe, from the state of their preservation, that they belong to the period 
of retreat of the ice sheet at all, and not to the period of its growth, we are forced to 
conclude that the advance which they indicate must have been of the nature of an 
oscillation affecting the general retreat. 

In applying this line of argument it is necessary, in the first place, to compare only 
valleys that are dry, so as to eliminate uncertainties due to post-glacial erosion. It will 
be noticed that we here compare the top of A and the bottom of B. Post-glacial 
change can only be supposed to act so as to lower the former, and, in the case of a dry 
valley, to raise the latter so that if the top of A be still higher than the bottom of B we 
may confidently believe that the difference was no less marked in glacial times. 

The oscillations of the East Lothian ice sheet have left several memorials of the 
type described above. Black Law, a little to the west of the kamiform gravels of 
Woodhall, furnishes an excellent example, for instance ; so also does the great Rammer 
Cleugh channel behind Deuchrie Dod, # described by Professor Young, but in the last- 
named case there is additional evidence showing that oscillations at this stage were 
repeated more than once. In the first place, a high-level channel, as shown on the 
general map, obviously furnished the original eastern continuation of this overflow 
channel, and must have been blocked by a re-advance of the ice before the eastern portion 
of the present course came to be employed. Then began the excavation of the magnifi- 
cent gorge which is known as Rammer Cleugh. It is two hundred feet deep and cut 
through solid rock, the finest example of a dry valley in the county. It is hardly 
conceivable that the great ice sheet retreated regularly and slowly during the cutting of 
this deep channel, since under such circumstances it only had a mile in all to travel 
before presenting the marginal drainage with a fresh path, for escape, lying to the north 
of the Dod. We cannot, however, point to any direct evidence indicating oscillations 
at this stage ; but later, when the cleugh had been excavated to its present depth, a 
re-advance is clearly indicated by a group of morainic mounds charged with great blocks 
of red sandstone carried from the lowlands, which rest actually on the bottom of the 
valley near its western extremity. It is interesting to note that while this line of 
moraines was breached by a continuance of the drainage in an easterly direction, the 
small stream which now occupies the gap flows in the reverse direction as a result of 
subsequent corrom formation. 

The same kind of evidence is rendered available owing to records of underground 

* Pl. I. % 1. 


operations in the neighbourhood of Ormiston, of which Dr Crampton has kindly given 
us the following interesting account. " To the west of Ormiston, the light railway 
follows what has at one time been an important line of drainage. It is shown by bores 
to be in large measure choked by alternating deposits of sands and gravels, and it is 
occasionally so completely filled up that its edges no longer make a feature at the 
surface. Another channel lying to the south and east of Ormiston, and passing close by 
Pencaitland, has also been proved to contain a considerable depth of clay and sand, and 
from some of the bore records it can be definitely recognised that boulder clay here 
occurs superimposed upon sand and mud. The journal of one of these reads as 
follows : — 

Surface, 1 foot. 

Clay and stones, 17 feet. 

Sand and gravel, 13 feet 10 inches. 

The hollow, despite its partial infilling, is still strongly marked and appears to have 
been reopened to some extent by the flow of water at the time of the retreat of the ice 
sheet which had deposited the boulder clay mentioned above. The explanation which 
seems most applicable to these two channels is that they were marginal stream courses 
formed in front of the ice sheet during its oscillatory retreat, and therefore liable at 
times to be overwhelmed by a temporary re-advance of the glacier and in this manner 
choked with boulder clay." 

III. The Upper Limit of Marginal Phenomena set by Local Glaciation. 

We have, up to the present, been dealing with the phenomena which marked the 
retreat of the great ice sheet from about the 1000-foot level downwards. Above this 
the processes at work were evidently somewhat different, for they have left a record of 
their operations which, though meagre, is still eloquent of the special conditions 
prevailing during the earlier phases of the retreat. As soon as the Lowland ice had no 
longer power to surmount the Lammermuirs, the dissected plateau in which this range 
culminates became for the time being an independent area of dispersal, though still 
retained within the confines of the great ice sheet. Even when the latter had shrunk 
back for some little distance from the summits of the hills, every hollow and valley was 
choked with neVe\ and many a one nourished its own little glacier, so that every here 
and there a tongue of ice still reached down the gentler slopes to join its great 
companion in the lower ground.* So long as this last condition held, it is obvious that 
marginal drainage on a large scale must have been impossible, and we have at once the 
explanation of the upper limit referred to above. Later, however, a stage arrived when 
the Lowland ice parted company with the local glaciers, for while the former continued 

* It must not be overlooked that the evidence adduced in this and the preceding section has » most important and 
direct bearing upon the climatic conditions of the period. 


its retreat from the hill country, the latter shrank back into further recesses of the high- 
level valleys and finally dwindled away to nothing. 

Very simple and straightforward evidence of local glaciation is afforded in the 
neighbourhood of both Lammer Law and Soutra Hill. The former case is illustrated in 
fig. 4. The Sting Bank Burn follows a deeply trenched valley running eastwards from 
Lammer Law. A neatly cut high-level channel leads right across the ridge bounding 
the valley on its northern side. It shows all the characteristics of a winding stream 
course, with steep banks facing every curve. Its elevation above the Sting Bank Burn 
is about 200 feet ; it is now quite dry, and as it cuts right through the ridge it has no 
catchment area at its head whatever. It follows the bottom of a faintly marked 
transverse col, which is lower than any other point for some distance along the ridge to 
the east. To explain this high-level channel we must admit that the course of the Sting 
Bank Burn was choked with ice streaming in from the tributaries to the south, but that 
a small lake was permitted to form in the north-eastern angle of the valley which 
drained across the col at its lowest point, and thus led to the cutting of a stream course in 
what one would now style an impossible position. As the water which cut this channel 
drained outwards from the hill country we have clear indication that local glaciers 
remained in the high valleys after the great ice sheet had shrunk to a lower level. We 
need only notice that the configuration of the land emphatically negatives any suspicion 
that the great ice sheet could block the mouth of the Sting Bank Burn or the Hopes 
Water into which the latter Hows, so as to form a lake whose outflow might have caused 
this channel, for under such circumstances there could have been no possible escape for 
a stream across the col, which would itself have been submerged. 

An exactly similar deserted high-level channel leading outwards from the hill 
country is shown on the general map just west of Soutra Hill (see general map) ; 
and while it confirms the deductions we have already drawn, it also enables us to 
appreciate an interesting case of stream diversion which has occurred upon the other side 
of the same hill. Originally, or rather in pre-glacial times, the Armet Water had its 
source at Nine Cairn Edge, for the valley still continues naturally in that direction, but 
the upper portion of the stream has now been diverted to the north and has given rise 
to the Linn Dean Water. It seems clear on the ground that the local glacier, which 
occupied the whole of the Armet valley, protruded a tongue northwards to the east of 
Soutra Hill. The high-level gravels on either side of the Linn Dean Water represent in 
a sense the morainic deposits of this tongue of ice, while the gorge of the stream was 
cut by the waters issuing from the same. Certain minor diversions of the stream, more- 
over, suggest that the glacier sometimes extended forward even into the water-worn 
channel of the burn and forced it to cut a new course for itself a little to the one side or 
the other. 

It is not impossible that the diversion of the Blinkbonny Burn (north of Sting Bank 
Burn) should also be ascribed to a precisely similar interference on the part of a local 
glacier. Originally this stream joined the Harelaw Burn to the east. Such examples 



illustrate the final stages which may be reached in a development originating in so small 
a thing as the high-level channel leading across the col out of the Sting Bank valley. 

In the cases so far considered there is no direct evidence * that the local glaciers 
of the higher valleys came into being during the retreat of the main ice sheet. The 
excellent preservation of the stream channels described proves indeed that they were 
produced at a later date than the final overwhelming of the hills by foreign ice, while 
the oscillatory nature of the withdrawal of the latter itself suggests that climatic 
conditions were favourable, at the time of the retreat, for a temporary retention of small 
glaciers in the valleys of the hill country. Fortunately, however, this important point 




glaciers glacial lames glac/al drainage chanhels hill summits watersheds scale of miles 

Fig. 4. 

can be settled definitely by a consideration of the complicated series of events recorded 
in connection with the local glacier of the Cowie Burn valley (figs. 4 and 4a). 

The critical sections occur along the ridge which runs northwards from Nine Cairn 
Edge and forms the western boundary of the Cowie Burn valley. Attention will be 
restricted to the system of channels lettered in the text map. Of these only the 
example marked A B C is significant in regard to the question of local glaciation, but it 
would be unsatisfactory to attempt to deal with it without explaining the origin of 
the others. 

The first thing that strikes the observer in visiting this ridge is that the channels of 
this system have delivered water (they are all now perfectly dry) both in an easterly and 

* We prefer, for the present, not to regard the absence of the ordinary marginal phenomena from the region of 
local glaciation as direct evidence on this question, to avoid reasoning in a circle. 


in a westerly direction. The phenomena are well marked and quite unmistakable. To 
understand this interesting complication one must recognise that the system of channels 
lies wholly above the 900-feet contour ; that is, above the summit level (884 feet) of the 
Heriot railway pass, six miles to the west. An examination of this pass shows that it 
was actually employed as a glacial drainage channel by waters travelling from north to 
south across the hills. As a matter of fact the pass was sometimes used in the reverse 
direction also, but when available for the northern drainage, it of course offered an 
opportunity of escape for waters empounded in front of the retreating ice sheet for some 
distance to the east of its entrance as well as to the west. 

Now a glance at the general map will show the even course followed by the 1000- 
feet and 900-feet contours between the mouth of the Heriot Pass and the northern 
slopes of Lammer Law, and it will be readily understood that very slight differential 
movements of the ice front might lead to flow of water in the one direction or the other 
across the ridge we are at present considering. 




cohiE burnCLA_ 




Fig. 4a. 

Having discussed this point, the next question is as to whether the channel ABC, 
draining east, was cut earlier or later than the channel BD which drains west by various 
mouths. The answer is attained by a line of argument already developed in a previous 
section. The top of the channel cut at A is higher than any point in the bottom of the 
channel B D draining west from B, therefore it would seem that A was cut before B D. 
The water entering # at the intake A flowed across the low col in the water shed at B, 
and then, instead of following the uniform slope of the hillside down into the Cowie 
Burn, was deflected so as to run along the side of the ridge, cutting a channel B C, forty 
feet deep, in this position. One can scarcely avoid the conclusion that this channel 
started in a gutter of which the south-eastern wall was the front of a small glacier 
occupying the Cowie Burn valley. Once, of course, the stream had lowered its bed into 
the rock of the hill side it would become self-supporting and could dispense with aid 
from the local glacier. All the time, however, the great ice sheet must have remained 
to impound the lake that discharged through A. We are justified, then, in picturing a 
time when these two glaciers, the large and the small, faced one another across the 

* This is the condition of affairs shown in fig. 4, although the other channels of the district are also shown. 


narrow ridge, and the contemporaneity of the local and foreign glaciation of the district 
can scarcely be doubted. 

The next stage in the history, so far as it can be unravelled, was marked by an 
advance of the great ice sheet blocking the intake A, and also the escape by way of the 
Lammer Loch channel to the east. At the same time the Heriot pass, far to the west, 
happened to be available, so that drainage took place along the ice front cutting the 
channel BDE. A partial retreat allowed the water to follow another course B D G, and 
further retreat allowed of lateral escape simultaneously from two points along this 
channel leading to the excavation of the cross contour courses D H and F I. That these 
two did come simultaneously into operation is almost certain from the fact that the bottom 
of the lateral channel leading out at D is slightly deeper than the high-level entrance to 
F G, so that had D H been formed first, F I could not have been reached by water, and 
therefore could not have been excavated at all. On the other hand, it is impossible to 
imagine F I cut first, for its exit could not have been closed again by the ice sheet 
without also closing that of D H. Further, there is direct evidence in the field that 
the original depressions at I) and F, before either was breached, must have been at 
approximately the bottom level of the groove B F G. Once these lateral escapes were 
opened and cascades had formed on the steep hill slopes, the channels were rapidly 
deepened. F I seems to have offered a freer escape at first, for it evidently carried the 
greater bulk of water. 

We did not make a particular study of the deposits of the local glaciers of the 
Lammermuirs. No doubt they supplied a considerable proportion of the material of the 
great Lammermuir terraces of sand and gravel, but we consider that the chief memorial 
of their existence is the protection which they provided to the upper part of the range 
from the ravages of marginal stream action. At the same time the following description 
by Sir Archibald Gkikie * seems to indicate that the local glaciation may have been 
responsible for the production of a certain amount of morainic material. We quote in 
full : " Between the heath and the rock there intervenes, on the higher parts of the 
Lammermuir range, a mass of rudely stratified rubbish, to which, in the progress of the 
survey, the provisional name of Surface Wash has been given till its true history is 
better understood. It' lies along the ridge between the head of the Hopes Water and 
the Kilpallet Heights. It is usually a sandy clay or earth more or less distinctly 
stratified, and containing pieces of greywacke and shale, often well striated. Though 
frequently different in character from the rocks immediately adjacent, these stones 
cannot have come from any great distance. The deposit is only a few feet in thickness. 
Perhaps it is to be regarded as the remanie of the boulder clay, washed down and 
reassorted by rain, though the comparatively small number of the striated stones and 
the irregularity of the striation have sometimes suggested that this accumulation may 
be as old as the snowfields of the glacial period." 

* Geology of East Lothian, 1806, p. 65. 


IV. The Influence of Corroms and Bridge Deltas in Modifying 
the Glacial Drainage System. 

We have already seen that the glacial drainage channels of East Lothian have 
characteristically a very gentle gradient. This feature owes its origin to the magnitude 
of the torrents which the channels were called upon to cope with during the intermittent 
floods of summer-time. Now it frequently came to pass that a low-grade valley of this 
type was established at one stage in the withdrawal of the ice sheet, to be left later on, 
as the retreat continued, open at its head and deprived of its main supply of water. 
Some little trickling stream might still perhaps wind its way along the bottom, but such 
would obviously prove quite incapable of coping with the tributes of sand, gravel and 
alluvium delivered by burns descending into the gorge on either hand. Deltaic cones 
have been formed at the entrance of every side stream.* They are sufficiently 
conspicuous to be easily recognisable, and resemble exactly well known instances in other 
localities. t In the upland districts they furnish sites for the shepherds' cottages, and 
include the only patches of land brought under cultivation. They not infrequently 
obstruct the insignificant drainage of the main valley, and so give rise to small tarns 
and marshy tracts situated in their rear. 

By way of illustration the number of cones which have been formed along the course 
of Edmonds Dean was determined (see general map and fig. 5), and attention was paid 
to the path traced out by each tributary stream in flowing down its delta. Thirteen 
cones occur in a distance of two miles, and of these, six drain centrally, two down 
stream and five up. The drainage directions here given refer only to what occurs on 
the actual surface of the cones, for on reaching the bottom of the main valley all the 
tributaries turn to flow south-east — the original direction followed by the glacial 
drainage. The cone occurring furthest up stream has, however, very nearly raised 
itself to the entrance level of the pass, and an inconsiderable addition to its height 
would develop it into a corrom or delta watershed. 

The number of fully-formed and operative corroms modifying the glacial drainage 
plan of the county is so great that an enumeration is impossible. To realise the 
important role which some of these corroms play, one has only to seek out the great 
highways which were established, under conditions of glacial restraint, to connect up 
drainage areas previously distinct. In almost every case the striking feature here 
is the very minor importance of the re-distribution of drainage permanently effected ; 
the lines of communication temporarily employed are now blocked by the formation 
of corroms which coincide roughly in position with the old pre-glacial watersheds. 
Further, the slight interchange involved in this process has taken place, more often than 
otherwise, in a direction contrary to that of the old glacial stream. The cause of this 

* PI. III. fig. 1. 

t J. E. Marr, Scientific Study of Scenery, second edition, p. 166 and pi. I. C. Russel, River Development, 
p. 138. 



is not far to seek. Suppose that a natural drainage basin A is forced by glacial 
obstruction to seek an outlet through a neighbouring basin B, and that a level-bottomed 
gorge is thus established connecting the two. On further retreat of the ice sheet a 
readier way of escape will likely present itself to the waters of the A basin, while those 
of the B basin may still persist along the same route to the sea^as heretofore. Under 
these circumstances it would be natural for a tributary of the B system to establish the 
corrom blocking the highway, since thus it would share in the advantages obtained by 
the A drainage. 

Four very notable illustrations are afforded in our district of the action of the 
processes which we have outlined above. Thus there is the striking instance of the 
railway pass between Cockburnspath * and Grant's House, which it will be remembered 
was specially noticed by Sir Archibald Geikie in his description ; then, again, there 
are the two great highways, in part occupied by the Aikengall and Thurston Burns 
respectively, to which Professor Young directed attention, and which serve to isolate 
in the one case Cocklaw Hill and in the other case Spott Hill from the main 
mass of the Lammermuirs. Last of all, far away to the west there is the Borthwick 
channel, which for so long a time connected the Mid Lothian and East Lothian 
drainage areas ; as mentioned before, Dr Crampton, working independently, has 
come to the same conclusion as ourselves with regard to the changes that have 
brought about the desertion of this last-named channel, and his account will shortly 
appear in the second edition of the Geological Survey Memoir dealing with the 
Edinburgh district. 

Let us now take the first of our four principal examples, that of the railway pass 
between Cockburnspath and Grant's House (fig. 5). The position of the pre-glacial water- 
shed, which formerly broke the continuity of this valley, can readily be recognised just 
at the point where the railway passes through a short tunnel to avoid taking an 
awkward little bend to the west. On the two sides of this old watershed the valley 
has passed through somewhat different stages of development. South of the col the 
glacial waters, for long supplied by the channel of Edmonds Dean, had nothing to do 
but deepen the old valley which served to conduct them towards the sea ; t but to the 
north of the col they had to deposit layer upon layer of gravel, sand, and silt, and thus 
choke the lake which formed in this position, while, when crossing the col itself, the 
waters were always busy cutting and digging. So it came to pass that depositing in 
one place and excavating in another, they never rested until they had graded the whole 
to one uniform gentle slope inclining from north-west to south-east. But now the ice 
retired and allowed the marginal drainage to escape round the north of the obstructing 
ridge by the St Helen's Church channel, described by Sir Archibald Geikie, and the 
great highway was abandoned which had cost so much labour to establish. The corrom 

* Pronounced Copperspeth. 

t A triangular mound near the southern exit of the tunnel is artificial, being composed of angular rock d^brify 
with occasional brick fragments, top-dressed with sand and boulder clay to allow of planting. 


now dividing the valley into two is a fairly conspicuous mound * built out by a little 
stream entering from Blackburnrig Wood ; it lies half a mile to the south-east of the 
pre-glacial watershed, and serves well to illustrate the general rule that a drainage area 
which has received water from a neighbour under conditions of glacial restraint has 
often to pay back with some of its own when that restraint is removed. The explana- 
tion given above clearly applies here, for the streams which continue along the route 
previously taken by the glacial torrent have not succeeded in deepening their course, 
while the Edmonds Dean Burn, which enters north of the corrom, but still south of the 

Fig. 5. 

pre-glacial watershed, has found a comparatively rapid drop by following the reverse 
direction, and has thus been enabled to cut a gorge for itself through the old lake 
deposits, which is fully a hundred feet in depth. A very conservative estimate of the 
advantage gained can be arrived at by a consideration of the following figures : the 
height of the corrom above sea-level is 396 feet, and this fall can be achieved in the 
one direction in three miles, as contrasted with twelve miles in the other. 

Turning next to the highway which leads round behind Cocklaw Hill (fig. 6), we find 
a similar instance of the removal of a pre-glacial watershed and the later formation of 
■a corrom further to the south. It is the Aikengall Burn which, entering the great 
through valley about a quarter of a mile beyond the site of the original watershed, has 

* PI. III. fig. 2. 



been the corrom builder here, and as a consequence it has transferred its allegiance from 
the southern to the northern drainage system. No valley can be more heartily 
recommended to anyone desirous of studying corrom formation than this one, owing to 
the number and beauty of the cones * which enter it rfrom either side. There seems 
indeed to have been some little competition as to whether the Aikengall Burn or its 
neighbour to the north should actually succeed in establishing the corrom of the valley ; 
but once the former had settled this point in its favour, it made sure of its conquest by 
attacking and partially demolishing the work of its rival. One is especially impressed, 


Fig. 6. 

both here and in the case of the great Dry Burn valley (Thurston valley) further north, 
with the probability that the main readjustments, which we can recognise to-day in the 
glacial drainage system, were virtually completed almost as soon as the opportunity for 
them arose through the continued retreat of the ice sheet. Thus, to judge from the 
erosion effects, it would seem that the reversed drainage, initiated upon the corroms of 
these two valleys, had formerly a much greater volume than at present, and one is led 
to speculate regarding a time when the streams were apt to be swollen into mighty 
torrents by the melting of the snows in the hill country. Although a corrom to be 
permanent must be built of sand, gravel, or some such material, we must not neglect 
the probability that the corroms of the glacial period may have largely resulted from 

* Pi. III. fig. l. 


the piling up of flood-borne ice and snow, mingled and covered with other debris, and 
that in many cases these temporary constructions may have had a very important 
influence in shaping the destinies of the drainage system. Before leaving this subject, 
we would like to notice a particular effect, for which the reversed drainage of the 
Aikengall Burn may fairly be held responsible. Just to the north of the entrance to 
the pass, perched upon its western bank, there occur two or three dry valleys, which 
there can be little doubt were used successively by the marginal stream at a time 
when the waters of the latter still found an escape southwards through the Aikengall 
valley. They now " hang" with respect to the latter, since, being streamless, they have 
not shared in the opportunity of renewed erosion, which it attained through the partial 
reversal of its drainage. 

After what has gone before, there would be no advantage in discussing the history 
of the Dryburn and Borthwick channels in detail. It suffices to point out that the 
streams responsible for corrom formation in these two cases are the How Burn and Gore 
Water respectively. 

We may now bring this part of the discussion to a close by observing that the 
operations of corrom formation are by no means restricted to glacial drainage systems. 
The circumstances which so greatly favour their growth in the latter are likely to 
be reduplicated in other situations. Wherever a gently graded valley lies open at its 
head and no longer supplied with a sufficiency of water, there a corrom is likely to 
appear ; and many a great beheaded valley must have found itself in just such plight, 
and accepted a corrom as the watershed of its lower reaches. The permanence of the 
capture thus involved can readily be predicted from analogy with the East Lothian 
types, for the diversion initiated upon alluvium would, as it were, be fixed by becoming 
imposed, through continued erosion, upon the rock framework of the country. 

It may be well to add here, since the term corrom is now being introduced for the 
first time into scientific literature, that the definition relates to function and not to 
origin ; it is on this account free from any speculative element. In the beautiful Glen 
Tarbert example (Ardgour), whence the name has been derived, the delta has accumu- 
lated on the surface of a rock watershed. A similar case has been described by Dr 
Marr * for Dunmail Raise in the English Lake District. The Glen Tarbert and 
Dunmail Raise deltas are corroms, since they serve as watersheds, quite irrespective of 
the fact that they cannot be held directly responsible for their behaviour in this 

While corrom formation is favoured in the major glacial channels, another modifica- 
tion has affected the less-marked grooves running along the hill slopes and roughly 
following the contours. Their fate can readily be pictured : the more important side 
streams have built out their cones into the channels, and thus have furnished themselves 
with bridge-deltas by which they have won their way to the other side. Once across, 
they have made good use of the slope and set to work eroding a channel for themselves, 

* hoc. cit., p. 143. 


often removing all evidence of the manner in which the crossing was made. A capital 
example of a bridge-delta occurs at Common House, rather more than two miles south- 
south-west from Spott. Several small streams uniting at this point have well nigh 
obliterated a portion of a strongly marked contour channel ; better still, however, they 
have later set to work again to destroy the delta, and in the deep gashes that they have 
cut one may recognise both the nature of the infilling material and the continuation, for 
some little distance, of the rock walls of the original gorge. 

While examples of bridging by deltas can seldom be demonstrated, it is still 
moderately certain that this operation is responsible in large measure for the reticulate 
pattern of the East Lothian glacial drainage system. 

There is, however, another and distinct way in which a glacial drainage channel may 
become segmented. Especially in the hilly districts, where pre-glacial features were 
never obliterated, the major streams have in many cases continued their course straight 
across glacial drainage channels, merely because the withdi'awal of the ice left an open 
path in this direction. 

It is obvious, then, that contour channels are very liable to division. When once 
this has been accomplished the next step is to provide each segment with its own 
independent watershed. Here of course an opportunity for corrom formation is 
afforded, and it is one that, in point of fact, is but seldom neglected. 

V. Evidence regarding the Level of the Sea at the time of the 

Retreat of the Ice Sheet. 

We have not made a special study of the intricate question of the relations of land 
and sea at the close of the glacial period, but attacking the question from a new stand- 
point certain results have been attained which we consider to be worth recording. 

In the first place, the phenomena which we have been describing are emphatically 
those of the retreat of an ice sheet obstructing land drainage, and become meaningless if 
viewed in the light of a submersion theory. In fact, to fix the upper limit which can be 
assigned to the level of the sea at the time of the retreat of the ice sheet we need only 
find the lower limit reached by the denudation effects of the land drainage when acting 
under conditions of glacial constraint. Caution must naturally be exercised in the 
choice of evidence upon which deductions of this character are to be based, and none 
but " dry valleys " should be employed for the purpose. All the necessary requirements 
were found in an exceedingly well marked glacial drainage channel cut in rock and 
terminating near Ewe ford, slightly more than a mile south-west from Dunbar ; the little 
water that flows through it merely trickles along a ditch which has been dug to drain 
the marshes which still occupy portions of its bottom. We may confidently assert that 
under natural conditions the valley would be streamless ; and it is obvious that so far 
from having been overdeepened since its desertion by the marginal stream, all the 
processes at work have tended to its obliteration. In spite of this, however, it only 


ceases to be a well marked feature on crossing the 75-foot contour line, so that 
high-water mark at Dunbar, when this channel was being cut, cannot have reached the 
level of the 100-foot contour line of the present time, and probably did not surpass the 
7 5 -foot contour. 

In the time at our disposal we were unable to follow up this promising line of 
evidence, but a visit to St Abb's showed a very similar relation. Lying in the bottom 
of the hollow which isolates the headland, is a well defined glacial drainage channel. 
It cuts through rock covered with boulder clay, and is clearly recognisable well nigh as 
far as the 50 -foot contour line. 

One further observation may be added which may some day prove of interest in 
dealing with the raised beaches of the Forth. About half a mile south of the railway at 
Prestonpans a clearly marked glacial drainage channel cuts through the 150-foot contour 
line. It in part follows the course of an old glacial groove, but at the same time it 
undoubtedly shows that the ice sheet at the time of its formation lay to the north and 
not the south. We see, then, that the great ice sheet continued its process of shrinking 
into the centre of the valley even to this late stage, and that the bed of the shallow 
estuary of the Forth served as its last refuge when it had withdrawn from the coastal 
regions of East Lothian. 

Trans. Boy. Soc. Edin. 

Vol. XLVI. 

Professor P. F. Kendall and Mr E. B. Bailey on " The Glaciation of East Lothian 
South of the Garleton Hills." — Plate I. 

Fig. 1. — Rammer Cleugli, lying south of Deuchrie Dod (the latter not included in the picture). This is the finest dry valley of the 
district : cut 200 feet deep in rock. Note high-level channel to north of the main valley (left side of the photograph), cut at an 
earlier stage and blocked during a re-advance of the ice, which led to its abandonment and the opening of the Rammer Oleugh 
channel proper (pp. 9, 19). 

Fig. 2. — Very typical dry valley joining the Spott Burn a short distance above Spott. Note the high-level channel to the south of 
the main valley (right side of photograph), and probably cut during a temporary re-advance of the ice sheet (p. 11). 

l\ans. Roy. Soc. Edin. 

Vol. XLVI. 

Professor P. F. Kendall and Mr E. B. Bailey on " The Glaciation of East Lothian 
South of the Garleton Hills."— Plate II. 

Fig. 1. — Junction of the dry valley shown in PI. I. fig. 2, with that which carries the Spott Burn. 
carrying the stream has not been appreciably deepened since glacial times (p. 11). 

In this case the valley 

Fig. 2. — Braidwood Burn. The present stream has cut a minor winding gorge into the bottom of the glacial channel, 
can still be readily recognised. The material cut through by the burn is Old Red conglomerate (p. 12). 

The latter 

>*AI_ ^ 

Trans. Roy. Soc. Edin. 

Vol. XLVI. 

Professor P. F. Kendall and Mr E. B. Bailey on "The Glaciation of East Lothian 
South of the Garleton Hills." — Plate III. 

Fig. 1. — Aikengall Valley. A delta is here shown, deposited by a tributary stream where it enters a dry valley. Note the 
difference of vegetation on the rock sides of the valley, on its ill-drained bottom, and on the slopes of the delta respectively, and 
see how the sheep cluster upon the latter (p. 25). 

Fig. 2.— The corrom of the Grant's House railway pass, deposited from the small wooded gorge of Blackburnrigg on the far side of 
valley. Note how the telegraph posts well nigh disappear in the cutting through the corrom and reappear on the further side 
(pp. 8, 27). 

Trans. Roy. Soc. Edin. 

Vol. XLVI. 

Professor P. F. Kendall and Mr E. B. Bailey on 
South of the Garleton Hills. "- 

: The Glaciation of East Lothian 
-Plate IV. 

Fig. 1. — Sands exposed by Keith Water, one mile north-west of Upper Keith. The top bed in the section consists of 6 feet of boulder 
clay, which forms the flat on which the railing stands. Note hummocky form of country, due to denudation. The section is, 
in descending order, boulder clay 6 feet, sand 65 feet, boulder clay 33 feet (pp. 14, 17). 

Fig. 2. — Sands exposed by Keith Water in the Red Scar, 1| miles west of Upper Keith. The bed of boulder clay which caps the 
section shown in fig. 1 is here overlain in turn by a thick mass of sand. The sands underlying this boulder clay are generally 
obscured beneath rain-wash, which has been in part cleared to allow of this photograph being taken. The intercalated boulder 
clay indicates an oscillatory re-advance of the ice sheet which served as the northern boundary of the lake in which the sands 
accumulated. The section is, in descending order, sand 53 feet, boulder clay 13 feet, sand 15 feet, boulder clay 37 feet (pp. 14, 17). 

Trans. Roy. Soc. Edin r 


•* Best sections showing intercalation of Sana's and Boulder Clays. '■ ' Sand and Gravel- 

Heavy solid lines = Dry Glacial Drainage Channels. Broken heavy lines = Stream. - containing GlM'1 %j t / 

Vol. XLVL 


M'FuOu< J; Etito, lut Edu£ 

( 33 ) 

II. — The Plant Remains in the Scottish Peat Mosses. By Francis J. Lewis, F.L.S., 
Lecturer in Geographical Botany, University of Liverpool. Communicated by 
Professor Geikie, LL.D., D.C.L., F.R.S. (With Four Plates.) 

The Scottish Highlands and the Shetland Islands. 

(MS. received June 25, 1907. Read July 8, 1907. Issued separately November 18, 1907.) 



Brief review of the work in Scotland . 

Method of field work 

The areas investigated — 

1. Valley deposits in the Highlands . 

2. Upland deposits in the Highlands 

3. The Outer Hebrides 

4. The Shetland Islands 

Previous work on Pleistocene plant deposits in 

Britain ........ 






Relation of the peat strata to other deposits in 

Britain 57 

The Pleistocene plant deposits of the Continent . 58 
The geographical distribution of the successive 

strata 59 

The relation of the strata to the glacial succes- 
sion 66 

Summary ........ 68 

List of References 69 

Explanation of Plates 69 

During 1904 and 1905 the author investigated the plant remains of the peat deposits 
in certain areas of the Southern Uplands and the Highlands. The present paper deals 
with similar work over large peat-covered areas in the E. and N. W. Highlands, Shetland 
Islands, and the Outer Hebrides. 

Perhaps the most interesting feature found in the course of the work is the evidence 
of constant change in the character of the vegetation which has successively covered 
the peat areas of Scotland. Many areas in different parts of the country have been 
examined in order to ascertain if this succession can be accounted for by local changes 
in physical conditions, or whether it represents stages in a long cycle of secular changes 
in climate, and consequent plant migrations. 

If the successive changes jn vegetation are constant over large areas in different 
parts of the country, some of the main facts for a distributional history of the British 
flora are available. 

The scope of such an inquiry is limited by the comparatively small flora characteristic 
of peat, but, on the other hand, the peat deposits possess certain advantages, for the strata 
are laid down in a regular series, and thus correlation of the strata in one district with 
those in another does not usually present any serious difficulty. 

As all the peat deposits are stratified, it becomes a matter of interest to examine the 

plant remains contained in the several strata. The search might be expected to yield 


evidence of interest both to botany and geology. On the one hand, it will give the facirs 
of the vegetation which has during successive periods covered the peat mosses ; and on 
the other, much may be learnt of the climatic conditions from the presence of certain 
sets of characteristic plants in the strata. 

The work of the last three years has shown that many of our peat mosses began 
their growth under arctic conditions, and thus, during some stage of the glacial period. 
The successive strata in such areas will then contain the salient features of the types 
of vegetation which have existed from that time down to the present. Direct palseonto- 
logical evidence will then be available for a long period of time during which, geologists 
are agreed, many climatic changes have occurred, though the amplitude of such changes 
is still a matter of controversy. Differences in climatic conditions of small amplitude 
would certainly affect the character and distribution of vegetation to a greater degree 
than the alluvial drifts which were deposited whilst those changes were in progress. 

Before giving a detailed description of the deposits examined during the field work 
in 1906 it may be well to review briefly the results obtained in 1904 and 1905. 

During the examination of large areas of peat it is found that certain datum lines 
make their appearance in all the districts. The most prominent datum lines are two 
beds of buried forest separated by a considerable thickness of intervening peat which is 
always free from tree remains. These two forest beds — the lower and the upper buried 
forest — are found in all large and deep peat mosses above 100 feet altitude which 
have been examined in the south of Scotland. The lower forest consists of birch, hazel, 
and alder ; the upper generally of pine. These forest remains are so well marked, and 
persist over so wide an area, and are underlaid and overlaid by such closely similar strata 
in the different districts, that there can be no doubt they represent true horizons. 

The upland valley deposits of the south of Scotland show a bed of peat containing 
the remains of an arctic flora between the lower and upper forest bed. This is considered 
to indicate a decided change in conditions between the dying away of the lower forest 
and the appearance of the upper forest. Those peat areas below 300 feet, and near the 
sea in Kirkcudbrightshire and Wigtownshire, contain beds of swamp vegetation between 
the two forest beds instead of arctic plants, but these swamp plants are such that attain 
a high latitude in Europe and Asia. In the south of Scotland the peat thus shows 
a transition from a forest vegetation to one characteristic of Arctic regions, and this 
gradually gives place to a second forest vegetation, which in turn is displaced by 
bog and moorland vegetation persisting to the present time. 

The Highland areas so far investigated reproduce the later stages found in the 
Southern Uplands, but the earlier beds are wanting. The basal layers of the peat 
contain the remains of an arctic vegetation similar to that met with between the two 
forest beds in the south. Above this, lies the upper peat bog covered with the upper 
forest and recent peat. The upper forest consists of two distinct zones separated 
by 1-3 feet of Sphagnum peat quite free from traces of wood. This is a feature of 
constant occurrence in the Highlands, but it has not been found in the south of Scotland. 


Field Work. — All the peat areas examined lay some distance from crofts or 
villages, and the strata showed no sign of having been disturbed by human agency. 
Sections were cut down through the successive strata until the underlying rock, gravel, 
or sand deposits were reached. Each section was made just large enough to work in 
conveniently. Over areas where the peat was of moderate depth the sections were 
usually 8 or 9 feet in length and 4 feet wide. In some districts, such as the Shetlands, 
where the peat proved to be 18-19 feet in depth the sections had to be made 17-20 
feet in length to allow a series of platforms to be left at each end, upon which the peat 
from near the base of the section could be placed. In some of the areas a good deal of 
difficulty was experienced in cutting through the upper forest of pine. The timber 
proved to be so large and tough, and so closely packed, that it was impossible to cut 
through it with a spade, and it was necessary to clear away the large trunks and roots of 
the pine with the aid of an axe and saw before the underlying peat could be excavated. 

In the examination of the peat samples brought back to the laboratory, much help 
has been obtained by the use of the method described by Gunnar Andersson (l). 
Samples of the peat are treated with nitric acid, which completely oxidises the material, 
loosens, and at the same time bleaches it. The peat is then placed in sieves and the 
loose debris washed away, leaving any seeds and fruits which it may contain. Fragile 
leaves are usually broken into fragments during this process, and they can best be 
obtained by dissecting them out from the peat without previous treatment. 

The following areas are described in this paper : — 

1. Valley deposits in the Highlands. 

(a) The Parph, Cape Wrath. 

(6) Assynt district of Sutherlandshire. 

2. Upland deposits in the Highlands. 

(a) Water of Lee district, Grampians. 

(b) Rannoch Muir. 

3. The Outer Hebrides. 

Eastern, Central and Western peat bogs of Lewis. 

4. The Shetland Islands. 

Walls, and central districts of the mainland. 

1. Valley Deposits in the Highlands. 

(a) The Parph, Cape Wrath (one-inch Ordnance Survey — sheets 113, 114). — 
This area is roughly triangular in shape, lying between Cape Wrath, the Kyle of 
Durness, and Loch Inchard. The northern part towards Cape Wrath consists of a 
series of somewhat rounded hills rising to 1500 feet with broad valleys between, covered 
with peat which extends far up the hillsides. Farther to the south the country becomes 
rugged, and some of the hills, such as Creag Riabhach, show very fine rock exposures, 
but peat still occurs in the valleys. 



A good bird's-eye view of nearly the whole district can be obtained from the summit 
of Fashven, six miles south of Cape Wrath. Seen from here, the peat deposits appear as 
a broad sheet covering all the lower part of the country, only the bare Torridonian 
sandstone summits being free from peat. In the more exposed situations denudation 
has been active, large patches of peat are bare of vegetation and frequently under water. 

The surface of the peat and the configuration of the peat-hags on the summit of 
Maovally bear witness that the action of the wind is a most important factor in the 
denudation of the peat. On the hill slopes, particularly those that face N.W., the 
rhizomes of Junceae all lie on little ridges of peat held together by the numerous 
fine roots of the plants, the peat having been completely blown away from the sides 
and tops of the rhizomes. The summit of Maovally (984 feet) forms a small plateau 
about one-third mile across. This was originally covered with deep peat, which now 
has been washed away to forms exactly resembling sand-dunes — the longer axes of the 
dunes all pointing N.W. to S.E., the direction from which the strongest gales occur. 

The peat over this area supports a vegetation of Scirpus ceespitosus * and stunted 
Calluna, forming aD association much like that described by Pethybridge and Praeger (2) 
on some of the Irish moors in Wicklow. No trees of any kind now occur in the district. 

The chief geological formations are Torridonian Pebbly Grits and Gneiss (Lewisian). 

Sections and borings were taken in all the valleys and up the hillsides for a distance 
of nine miles south of Cape Wrath. Owing to denudation, natural sections are fairly 
numerous, particularly in the north. The peat was found to be of moderate depth, 
varying from 8-14 feet, the depth gradually increasing as some of the broad open valleys 
near Cape Wrath were examined. 

A preliminary examination of the peat banks throughout the district showed no 
trace of the upper pine forest except at one or two places to the east of Cape Wrath, 
where a few stunted pine trees were observed, but these were not comparable in size to 
those found elsewhere in the upper forest. 

The drainage system of the district is roughly divided into an eastern and western 
area by a succession of hills running south from Cape Wrath. 

Three typical sections may be taken from the western area. 

Section I. : — 

Characteristic Plants. 

1 . Scirpus cxxpitosus. 

2. Calluna vulgaris. 

3. Betula alba — small shrubs. 

4. Callvna vulgaris. 

5. Sphagnum. 
G. Km />■ I rum nigrum. 

7. Saliz reticulata. 

8. Dry hard peat mixed with fine angular sand ; a few 

small fragments of stems, but no other recognis- 
able remains. 


* The nomenclature of Hooker's Student's British Flora, third edition, is followed throughout. 



Accompanying Plants. 

Calluna vulgaris. 
Eriophorum angustifolium. 

5. Eriophorum vaginatum. 

6. Sphagnum. 

7. Potentilla Salisburgensis. 



Section II. 

Characteristic Plants. 

1. Scirpus caespitosus. 

2. Betula alba — small shrubs. 

3. Betula nana. 

4. Salix Arbuscula. 

5. Potentilla Comarum. 

Accompanying Plants. 

1. Eriophorum angustifolium. 
Calhma vulgaris. 

2. Eriophorum sp. (abundant). 

3. Empetrum nigrum. 

4. Empetrum nigrum. 

5. Oar ex ampullacea. 
Peaty sand and pebbles. 


Section III. : — 

Characteristic Plants. 

1. Scirpus, Sphagnum. 

2. Betula alba (small). 

3. Empetrum nigrum. 

4. Empetrum nigrum. 

5. Potentilla Comarum. 

Accompanying Plants. 
Calluna (fairly plentiful). 

Salix Arbuscula. 

Salix Arbusculu, Carex sp. 

Peaty sand. 

The sequence of strata in these sections agrees closely. The upper forest is every- 
where missing, but a constant feature is the presence of a shrubby growth of birch 
which in some cases extends from the Empetrum and Salix beds up to the recent 
Scirpus, Eriophorum, and Sphagnum bed. The birch is always of very small size and 
mixed with much Eriophorum, Calluna, and Sphagnum. In fact, the evidence points 
to the birch growth representing a period of peat formation or peat bog conditions, and 
differs very much in this respect from the features found in the lower and the upper 
forest. It may be recalled that a similar scrub birch mixed with peat bog plants was 
recorded from many other districts in the Highlands in L906 (3). In those districts, 
however, the upper forest of pine was well marked and occurred above the scrub birch and 
separated from it by some thickness of Sphagnum. Before fully discussing the sequence 
in this district, typical sections may be given from the eastern drainage area. 

Section I. : — 

Characteristic Plants. 

1. Scirpus peat. 

2. Calluna peat. 

3. Betula alba (scrub). 

4. Eriophorum vaginatum. 

5. Empetrum nigrum,. 

6. Potentilla Comarum. 

7. Structureless peat. 

Accompanying Plants. 



3. Eriophorum vaginatum, Sphagnum. 

4. Empetrum nigrum. 

5. Salix Arbuscula, Sagina nodosa. 


Section II. : — 

Characteristic Plants. 

1. Scirpus. 

2. Betula alba. 

3. Empetrum nigrum. 

4. Salix Arbuscula. 

Accompanying Plants. 

3. Menyanthes trifoliata. 

4. Empetrum nigrum. 


Section III. : — 

Characteristic Plants. 

1. Scirpus c&spitosus. 

2. Stiff structureless peat, having apparently been 

much exposed to denudation before being 
covered by the overlying Scirpus. 

3. Calluna peat. 

4. Betula alba (scrub). 

5. Empetrum nigrum. 

Accompanying Plants. 

1. Sphagnum. 



5. Salix Arbuscula. 

Peaty sand and rock. 

An interesting feature here is the frequent presence of peat underlying the 
Empetrum and Salix Arbuscula and S. reticulata zones. An examination of this basal 
layer in the field showed no recognisable plant remains, but samples tested with nitric 
acid in the laboratory yielded seeds of Potamogeton natans, P. preelongus, P. 
rufescens, Menyanthes trifoliata, Ranunculus repens, R. fiammula — in fact, the remains 
of an aquatic vegetation. All these species have a wide distribution at the present 
time, but their presence at the base of the peat suggests that the summers of that time 
were fairly warm and were long enough to allow these aquatic plants to seed freely. 
The shallow character of this bed and its absence in many of the sections rather points 
to its belonging to the same stage as the overlying Salix and Empetrum beds. The 
same feature has been met with underlying the arctic bed in the Shetlands, and its 
significance will be discussed more fully when dealing with that area. 

The general succession of vegetation here compares very closely with the succession 
in the Inverness-shire, Easter Eoss, and Caithness-shire areas described in 1906 (3). An 
important difference is the absence of the upper forest bed. This zone is generally 
present in the south of Scotland and in the Highlands ; it has been noted as far north 
as the valley of the river Dionard immediately to the south of this district, and is 
occasionally present in the peat on the moors round Loch Eriboll, but it has not been 
found in any of the peat areas examined in the Hebrides and the Shetlands. During 
the examination of the peat on the Caithness- Sutherland border the upper forest zone 
is generally found to be present but the pine is frequently replaced by birch. The 
northern boundary of this forest bed is evidently approached in the Cape AVrath district, 
as here, pine remains tend to become stunted and birch frequently takes its place. In 
most districts the upper forest bed dies out before the northern coast is reached. 

The scrub birch described from this district must not be confused with the lower or 
the upper forest bed, for the wood in all cases is very small, not plentiful, and is always 
mixed with quantities of Eriophorum, Sphagnum, and Scirpus remains. Moreover, 
it does not correspond in position to either the lower or upper forest bed. The 
sections show that the basal Salix and Empetrum beds were gradually replaced by peat 
bog associations, amongst which scrub birch managed to find a footing. 

As the peat deepened, the birch disappeared and Eriophorum and Sphagnum 
became dominant over the area. That conditions at this period were not suitable for 
the growth of birch forest is shown by the fact that the birch remains are often absent 


in the more exposed situations ; are always larger and persist for a greater depth in 
the peat lying in valleys which face to the south and east ; whilst on the hill slopes and 
in valleys which run north and north-west the remains are either very small or else 
entirely absent. 

This is in great contrast to the distribution of the lower or the upper forest beds, 
which are often found in positions now unfavourable for tree growth, such as the bleak 
and treeless Outer Hebrides, and at altitudes of 3000 feet in the Highlands. 

(b) The Assynt District of Sutherland. — This district is dominated on the east by 
the massif of Breabag, Conaveall, Ben More Assynt and Glasven, which rise to an 
average altitude of 3000 feet. An extensive region of undulating moorland lies to the 
south and west with Suilven, Canisp, Cul Mor, and Cul Beag rising to about 2600 feet, 
at intervals of some miles, as isolated and precipitous hills. The general elevation of 
the moorland is from 100 feet to 700 feet, and most of it is covered with deep peat, 
but the hills are bare and rocky. The peat is very little broken up by denudation, 
and the moors, intersected by scores of lochs, present a smooth and flowing aspect. 

Two areas were examined : 1, peat lying on quartzite in the valley of the river 
Loanan, to the south of Loch Assynt ; and 2, an area near Loch Urigill about nine 
miles farther south than the first area. 

1. Peat in the Loanan Valley. — The peat lies 300 feet above O.D. between Canisp 
on the west and the limestone cliffs of Stonechrubie on the east. The chief vegetation 
now is Calluna, Scirpus csespitosus, Eriophorum vagiyiatum, Narthecium ossifragum. 

The peat varies in depth from 5-8 feet ; the surface is but little denuded, and closely 
covered with vegetation. 

Many sections taken over the southern and higher part of the moss showed the 
following strata : — 

Characteristic Plants. 

1. Scirpus csespitosus. 

2. Pinus sijlcestris. 

3. Sphagnum. 

4. Betula nana. 

5. Betula alba of large size. 

6. Moss peat. 

Accompanying Plants. 

1. Eriophorum vaginatum. 

2. Calluna vulgaris. 

3. Traces of scrub, Betula alba. 

4. Empetnim nigrum. 

5. Corylus Aoellana. 

6. Calluna vulgaris (abundant). 

Stone pavement. 

This sequence shows three points of interest. Firstly, there are no arctic plants at 
the base of the peat. We may infer from this, either that the record of the earlier 
post-glacial vegetation is missing, or that the moss peat represents swamp conditions 
which prevailed at this place immediately after the retreat of the latest glaciers. The 
abundance of Equisetum remains in the basal zone in some of the sections taken in this 
area rather points to the latter view as being more probable. 

The second interesting feature is the presence of a bed of Betula alba (No. 5) and 
the Pinus sylvestris (No. 2) zone. The two layers of large trees separated by about 
1-1-| feet of peat closely correspond to the features observed in many Highland districts. 



In position the layer of Pinus (No. 2) and Betula alba (No. 5) agrees exactly with the 
upper and the lower layer of pine separated by 18 inches of Sphagnum peat recorded 
from the Spey-Findhorn watershed, Coire Bog, and the Findhorn- Nairn watershed (3). 
Here, however, the lower layer of trees consists of Betula alba, and there are also traces 
of this tree in the Pinus sylvestris zone. The occurrence of abundant Betula nana 
between two such forest beds is of some interest, as similar plants in the same relative 
positions have been observed in the mosses of the Grampian mountains and the 
Caithness-Sutherland border. 

The third noteworthy feature is the fact that Betula alba tends to replace Pinus, 
sylvestris. The same feature has been noted near the northern boundary of the upper 
pine forest. 

The northern and lower-lying peat of the moss yielded much the same sequence, 
although, as one would expect, some of the details differ, as the following section 
shows : — 

Characteristic Plants. 


Scirpus peat. 

1. Sphagnum. 


Pinus sylvestris. 

2. Calluna vulgaris. 


Scirpus ceespitosus. 



Betula alba of large size. 

4. Corylus Avellana. 


Rliacomitrium lanuginosum. 

5. Stunted Calluna, Scirpus 


Sandy peat with much Equisetum sp. 

6. Hypnum sps. 

Sand ai 

id rock. 

Accompanying Plants. 

It will be noticed that here the whole of the peat between the two forest beds is 
formed from Scirpus without any trace of Betula nana. 

2. Peat in the neighbourhood of Loch Urigill. — The peat occurs upon open undu- 
lating moorland to the west of the Ben More Assynt, and Breabag ridges. The mosses 
are generally level, and peat appears to be forming at the present time. Sections were 
taken round Loch Urigill and Crom Allt — a stream draining into the southern end of 
Loch Urigill. 

The upper forest zone is poorly represented, and pine is generally replaced by birch. 
Moreover, none of the sections showed any trace of arctic plants at the base, thus 
indicating that the earlier stages are missing. 

The general sequence is as follows : — 

Characteristic Plants. 

1. Scirpus ceespitosus. 

2. Pinus sylvestris. 

3. Scirpus rsesfritosus. 

4. Betula alba. 

Accompanying Plants. 

1. Calluna, Sphagnum. 

2. Calluna. 

3. Betula nana, Sphagnum. 
Rhacomitrium lanuginosum. 

4. Corylus Avellana. 
Sali.x purpurea. 

The Pinus sylvestris bed is absent in many of the sections, but the Betula, alba 
zone (No. 4) is generally well marked. 

It should be pointed out that the peat is continuous for many miles to the south 


and west, and that last year it was possible to examine only a comparatively small area, 
so that arctic vegetation may be present in other parts of the district. The sections are, 
however, of interest, as they confirm the widespread character of the upper forest beds, 
and also give evidence that this district is on the confines of the western boundary of 
the upper forest. 

2. Upland Deposits in the Highlands. 

(a) The Forfarshire Grampians (one-inch Ordnance Survey — sheet 65). — This area 
shows a great extent of peat, covering both the summit plateaux and sides of the 

The most characteristic plant occurring at present on the peat areas is Scirpus 
ceespitosus, and the general character of the vegetation is somewhat similar to that 
in Skye. Mixed with Scirpus csespitosus are Calluna in small stunted patches, 
Eriophorum vaginatum (scanty), Sphagnum, Cladonia rangiferina (abundant), and in 
the drier places Juncus squarrosus. Many of the flat summit plateaux have probably 
been the site of lochs which are now filled with peat. The vegetation in such positions 
is somewhat different, the chief plants being Rhacomitrium lanuginosum, forming large 
beehive-shaped mounds, Sphagnum, Eriophorum vaginatum, E. latifolium, Scirpus 
csespitosus, occasional stunted plants of Erica Tetralix, Calluna vulgaris, and much 
Drosera intermedia, with a few plants of Pinguicula vulgaris. 

The whole area is under deer forest. Sections were made on the summit plateaux 
round the head of Glen Lee, Glen Mark, and Unach Water, the altitude varying from 
2000 feet to 2700 feet. 

Three types of mosses were found in different areas in this district, and sections are 
given to illustrate each type. 

I. In the corrie between Drumhilt and Hunt Hill at 2000 feet: — 

Characteristic Plants. Accompanying Plants. 

1. Scirpus csespitosus. 

2. Pinus sylvestris. 

3. Car ex sps., Salix Arbuscula. 

4. Betula alba. 

5. Mossy peat, with traces of Betula alba. 

1. Eriophorum vaginatum. 

2. Calluna vulgaris. 

3. Betula nana. 

4. Calluna vulgaris. 

5. Viola palustris. 

Sand and gravel. 
II. On the moors round Unach Water at 2500 feet :- 

Characteristic Plants. Accompanying Plants. 

1. Scirpus csespitosus (abundant). 



1. Eriophorum vaginatum. 

2. Calluna vulgaris. 

3. Structureless peat. 

4. Betula alba. 

Sand and gneissose rock. 

4. Calluna vulgaris. 


III. Near Slidderies Burn at 2250 feet : — 

Characteristic Plants. 

1. Kriophorum vaginatum. 

2. Calluna vulgaris. 

3. Potentilla Comarum. 

4. B'-tnla alha (large). 

5. Salix reticulata, S. Arbuscula. 

Accompanying Plants. 

1. Calluna vulgaris. 

3. Salix Arbuscula. 

4. Viola palustris, Ajuga reptans. 

5. Equisetum sps., Carex sps., Sedum Rhodiola. 
Boulder clay. 

These three types are instructive, inasmuch as they show that peat of very different 
age may be encountered in dealing with a large district. 

In Section III. arctic vegetation occurs at the base of the peat resting upon boulder 
clay, but this is not the case in Sections I. and II., which are evidently younger than the 
third type. 

In Section III. Betula alba with Viola palustris and Ajuga reptans overlies the 
basal arctic vegetation ; in Sections I. and II. a similar vegetation occurs at the base 
of the peat. 

The remaining beds in the three sections show remarkable agreement in their main 
features, but they exhibit particularly interesting features in connection with the upper 
forest. The upper forest is, in this district, as in so many others in the Highland areas, 
in two distinct zones separated by about 2 feet of peat. The upper tier in Section I. is 
composed of pine and the lower of birch. The peat between shows a very different kind 
of vegetation , such plants as Salix Arbuscula, Potentilla Comarum, Betula nana, Carex 
sps., replacing the birch and pine. The intercalation of such arctic or sub-arctic plants 
between the two layers of the upper forest is not confined to the Grampians, as it also 
occurs in Caithness and Sutherland. The lower tier of forest forms a good datum line, 
as it is present in all three types of mosses. The intercalated arctic plants above are 
also present in the first and third type, whilst in the second a zone of structureless peat 
occurs at this horizon. In the second and third type the upper tier of forest is wanting, 
but in both cases a bed formed almost entirely of Calluna stems occurs at the same level 
in the peat. 

The presence of intercalated arctic or sub-arctic vegetation seems to give strong support 
to the view that there was a change of conditions between the growth of the lower and 
upper tier of the last forest period. At the present clay such plants as Salix Arbuscula, 
Betula nana, do not form moors at the same elevation and under the same climatic 
conditions as a forest of pine or birch. The observations of W. G. Smith (4) upon areas 
bordering the one under consideration show that the birch forest and accompanying 
vegetation extends up to about 1500-1750 feet, whilst the highest moors at 2000-3000 
feet are dominated by an association formed of Calluna and Vaccinium M:urtillus with 
Cornus Suecica, Rubus Chamsemorus, Betula nana, Lycopodium annotinum, Loise- 
leuria procumbent. In addition to this it should be remembered that most of the woods 
occurring so high as 1750 feet are of small extent and have clearly been planted, and all 
of them are protected either by dry stone walls or by some natural feature ; and as far 


as I know pine or birch woodland does not occur on any 2000-feet summit in the 
Highlands. In short, the distribution of pine and birch forest and sub-arctic moorland 
is well defined, and the two associations are not now found intermixed at the same 
elevations and under precisely the same climatic conditions. 

This alternation of forest, arctic moorland and forest, is either good evidence of a 
change of climatic conditions, or we must agree that in past times the different types 
of vegetation, for some unknown cause, bore altogether different relations to one another. 
The latter view can hardly have much to recommend it, and if accepted, would negative 
all stratigraphical evidence. 

We see, then, that whilst some types of the Eastern Grampian mosses began their 
history under arctic conditions, other types began to form during the latter stages of 
the upper forest period. In each type the younger strata are the same. Arctic plants 
occur between the two zones of the upper forest in several other areas in the Highlands, 
whilst in nearly all, the upper forest is divided into two zones, separated by some 
depth of peat indicating wet conditions. The presence of forest at such altitudes also 
indicates a different altitudinal range to that of the present day, but this will be 
discussed when describing the geographical distribution of the zones. 

(b) The eastern Peat of Ra,nnoch Muir (Ordnance Survey — sheet 54). — The moor of 
Rannoch forms an extensive undulating region lying at 900-1500 feet, bounded on the 
west and south by mountains rising to 3000-4000 feet, on the north by a chain of hills 
rising to over 2000 feet. The drainage from these great watersheds passes westward 
across the moor to Loch Rannoch, and thence by the river Tummel to Central Perthshire. 

The greater part of the moor is covered with deep peat, and this is particularly the 
case in the area south of Loch Lydoch and along the Allt Lochain Ghaineamhaich — a 
stream draining into the western arm of Loch Lydoch. In some places the peaty 
covering is interrupted by beautiful sets of moraines, whose steep sides are often covered 
with thin peat formed chiefly of Calluna. Observations were confined to the eastern 
part of the moor, to the south and west of Loch Lydoch — an area extending to about 
six miles by three miles. 

The vegetation over the peat-covered areas is dominated by Scirpus caespitosus, with 
stunted Calluna, Eriophorum vagmatum (generally distributed), Rhacomitrium lanu- 
ginosum (very abundant), Sphagnum (abundant on many of the spongy bogs near stream 
heads, particularly in the western part of the moor). The peat is now much denuded, 
and waste appears to be more rapid than growth. This is seen by the frequent exposure 
of the upper forest, although, where denudation has not been so active, it is buried 
beneath 4 feet of peat. Some difficulty was experienced owing to the amount of water 
in the peat, and many sections had to be abandoned from this cause. The general 
depth of the mosses ranged from 8-14 feet. 

Some interest attaches to the examination of an area of this description, as owing to 
the probability of the lower-lying or badly-drained portions being frequently flooded, 
much less regularity might be expected in the plant strata than on rounded hills and in 



broad, well-drained valleys. This expectation was fully borne out, but such differences 
were chiefly confined to the lower layers of peat. The upper forest is present over all 
the area examined and forms an excellent datum line (Plate IV. fig. 11). Sections taken 
to the S.E. of Loch Lydoch showed the following beds : — 

Characteristic Plants. 

1. Scirpus csespitosus. 

2. Pmus sylvestris. 

3. Sphagnum. 

i. Betula alba (shrubby growth). 

5. Structureless peat containing traces of Sphagnum. 

Accompanying Plants. 

1. Sphagnum. 

2. Calluna vulgaris. 

V actinium Myrtillus. 

3. Betula alba twigs. 

4. Eriophorum vaginatum. Sphagnum. 

Granitic saud and gravel. 

Between Loch Lydoch and the Black Corries the lower layers of the peat show a 
different vegetation : — 

Characteristic Plants. 

1. Sphagnum. 

2. Pinus sylvestris. 

2. Mossy peat with much Phragmites communis. 
4. Betula nana. 

Accompanying Plants. 

1. Scirpus csespitosus. 

2. Calluna vulgaris. 

4. Sphagnum, Equisetum sp. 
Potentilla Comarum. 

Further west, towards Kingshouse, the general sequence remains the same, but the 
basal beds yield a few other plants. 

Characteristic Plants. 

Scirpus csespitosus. 
Pinus sylvestris. 
Scirpus csespitosus. 
Grimmia sp. 
Betula nana. 

6. Equisetum, Sphagnum. 

Accompanying Plants. 

1. Sphagnum, Eriophorum vaginatum. 

2. Calluna vulgaris. 

3. Sphagnum, Eriophorum. 

5. Salix Arbuscula, Empetrum nigrum, Arcio- 
staphylos alpina. 

6. Many broken and water-worn fragments of Betula 

Sand and clay, with angular stones. 

It is evident from the plant remains that much of the moor of Rannoch peat began to 
form under marsh conditions. Quantities of water-worn birch fragments occur towards 
Kingshouse, not only with the marsh plants at the base of the peat, but in the drift 
below. The wood bears traces of prolonged water action, and although many of the 
fragments are small twigs, some are larger and evidently belong to fairly large trunks 
of birch mixed with fragments of birch bark. This wood evidently did not grow 
in situ, and it can hardly represent the drift of a few streams, as the material occurs in 
many sections spread over a large area. At the same time, this drift-wood cannot 
represent debris from the upper forest, as that is a well-marked zone in the higher 
layers of peat and separated from the drift-wood under discussion by strata-containing 
arctic plants. The character and position of this material certainly suggests derivation 
from older beds which existed here before the present peat mosses began to form. 


The peat immediately above the basal marsh vegetation forms a well-marked 
horizon, as it is entirely made up of the remains of Betula nana, stems and leaves, and 
scales of the female catkins, and Salix Arbuscula, Empetrum nigrum, and Arcto- 
staphylos alpina. None of these plants are characteristic of swamps or marshes, but 
Betula nana occurs in peat bogs at fairly high elevations, and the other plants are 
abundant on mountain slopes and summits. Comparing the Betula nana bed and the 
Pinus sylvestris bed (Nos. 5 and 2) in the last section, it is evident from the plants 
that the peat of these two beds was not laid down under the same conditions. No 
changes in the drainage or the character of the peat would cause a forest of Pinus 
sylvestris to replace a close growth of Betula nana, Salix Arbuscula, Empetrum 
nigrum, and Arctostaphylos alpina. At the present day these two types of vegetation 
have their own altitudinal range, and are never found associated. In the case of the arctic 
bed (No. 5) we are dealing with a depression of the arctic zone. The Pinus sylvestris 
bed represents a return of the forest zone to at least its present altitudinal limits. 

The upper forest bed (zone No. 2) does not apparently show the double tier of trees 
so characteristic in most of the Highland areas. The only indication of such a feature 
is in Section I., where Pinus sylvestris is underlaid, first by a Sphagnum bed and 
then by a layer of birch. But the wood is small in size, it is absent in many sections ; 
and where present it is always mingled with quantities of Eriophorum and Sphagnum ; 
a condition of things very different from that of a true forest bed, where the peat 
contains dry moorland plants and shows unmistakable signs of having been accumu- 
lated slowly. On the other hand, this shrubby birch agrees in character and in position 
with that described from the Spey-Findhorn and Findhorn-Nairn watersheds, Coire 
Bog, and Caithness in 1906 (3). The general sequence of the peat in this area agrees 
very closely with that in the areas just mentioned. 

3. The Outer Hebrides. 

Peat in the Island of Leivis (one-inch Ordnance Survey — sheets 99, 105). — 
Following on the examination in 1905 of the peat in North Uist and Skye, it was 
decided to examine part of the extensive peat in the Lews to see what light it would 
throw upon problems raised by the peat in Skye and Uist. In addition, the distance 
of the island from the mainland and its position on the extreme N.W. edge of the 
Continental plateau made it desirable to ascertain if the strata represented on the 
mainland also appeared there. 

The whole of the island north of Loch Erisort is practically peat-covered, with the 
exception of small cultivated areas near the townships on the coast. 

For a distance of 32 miles northward of Loch Erisort the island does not rise above 
900 feet, and most of it is under 400 feet. The central portion is almost level, and 
forms an unbroken peat bog over which are scattered scores of lochs. In the southern 
region towards Loch Erisort they are so numerous that the island consists of almost 


equal portions of land and water. Whilst many of the lochs are deep with rocky 
basins, others are merely hollows in the deep peat. The condition of the peat over this 
great area (where the conditions for peat formation are perhaps more favourable than 
in any other part of Great Britain) entirely confirms the conclusions reached by Gkikie 
(5) many years ago, as to the present rapid denudation going on in our peat bogs. 
Here is an area on the confines of the European plateau, exposed to Atlantic conditions 
and therefore having a moderate temperature, a heavy rainfall, and a moist climate — 
all conditions that are favourable for the growth of peat. In this area we find peat 
covering the floors of the valleys and the hillsides to a considerable depth, yet at the 
present time the peat is wasting away much faster than it is being formed. A traverse 
taken across the island from Barvas to Monung (the solitary hill rising to 800 feet in 
the north of the island), and from Monung to Stornoway, shows that for the greater 
part of the way the peat is cut up into a system of vegetation-covered ridges or banks 
separated by channels about 4 feet broad and 3 to 4 feet deep. Further, the wastage 
is, as we should expect, more noticeable on rising ground and hillsides, than on the 
levels, and is more marked as the eastern seaboard of the island is approached than on 
the west. Although this feature could be explained in some particular mosses by the 
raising of the surface, due to the activity of the peat-forming plants, the raised bog 
permitting freer drainage and thus paving the way for the succession of less actively 
peat-forming plants, such as the heather and hill-pasture associations, yet the feature 
is so universally present on all types of peat-bogs, on hill-tops, hill- slopes, and in 
valleys both in England, Scotland, and Ireland, that the ultimate cause must be sought 
beyond any such local and episodal events such as I have suggested. It is true that 
the present denuded condition of the peat is due to a change of vegetation, but the 
general replacement of Sphagnum associations and marsh vegetation by Calluna 
associations must be due to some cause operating over the whole of Britain, such as a 
general decrease in the humidity of the climate. 

The genera] vegetation of the peat mosses shows a close resemblance to those on 
the N.W. districts of the mainland. Calluna is fairly abundant, mixed with Scirpus 
casspitosus, Rhacomitrium lanuginosuyn, stunted Vaccinium Myrtillus, Potentilla 
reptans, Erica Tetralix, Narthecium ossifragum. 

Sections were made in three different areas in the Lews : — 

(a) In the centre and east of the island, to the S. and S.W. of Monung. 

(b) South of Barvas, and also in the basin of the river Bragor. 

(c) Submerged peat in Sandwick Bay, near Stornoway. 

(a) Peat Mosses South and South- West of Monung. — These lie at a general 
elevation of 200-400 feet, and form gently undulating moorland in the drainage system 
of the Gress river. The area contains many small lochs, some of which occupy hollows 
in the peat. 

In some places the remains of an arctic type of vegetation occurs at the base of the 



peat, overlaid by birch, and in other places the birch itself occurs at the base of the peat, 
the older beds being wanting. Peat-formation would hardly begin simultaneously over 
so large an area as this, and we should expect to find some areas dating back farther 
than others. Details of two sections will serve to illustrate this point and to give the 
general sequence over this area : — 

Section I. : — 

Characteristic Plants. 

1. Scirpus csespitosus. 

2. Betula alba. 

3. Erica cinera. 

Section II. 

Accompanying Plants. 

1. Sphagnum. 

2. Corylus A vellana. 
Alnus glutinosa. 

3. Emjpetrum nigrum (traces). 
Eriophorum vaginatum (scarce). 

Fine sand and rock. 

Characteristic Plants. 

Accompanying Plants. 

1. Scirpus csespitosus. 

2. Calluna vulgaris. 

3. Sphagnum. 

4. Betula alba. 

5. Salix Arbuscula. 

1. Sphagnum. 

2. - 

3. Scirpus csespitosus. 

4. Scirpus ccespitosus. 

5. Empetrum nigrum, 
Viola palustris. 
Carex sp. 

The younger layers of peat resemble those in Skye, but the basal beds of Salix 
Arbuscula, Empetrum and Carices, carry the history a stage further back. Section II. 
shows the more general sequence, where a thick bed of creeping willow occurs at the 
base of the peat. This type of vegetation does not now form a continuous sheet in any 
of the Outer Hebrides, or at similar altitudes anywhere in Britain, being confined to 
mountainous regions. The Betula alba zone consists of fairly large trees, mixed in 
some places with Corylus Avellana and Alnus glutinosa. The persistence of this zone 
in nearly all the sections, combined with other features to be described later, suggests 
that it represents the lower forest, and this correlation is supported by evidence from 
Shetland. The upper layers of peat in this district contain no trace of forest, and in 
this respect they agree with the mosses in Skye and N. Uist. 

(b) Peat South of Barvas and the Bragor River Basin. — The peat along the west 
coast has been much dug for fuel by the inhabitants of the numerous crofting villages. 
In some places the peat runs almost to the coast, in others there is a narrow fringe of 
poorly-cultivated ground a mile or so in width, protected from the sea in the neighbour- 
hood of Barvas by extensive sand-dunes. Eound the crofting villages of Barvas, Arnol, 
and Bragor, the peat has been entirely carried away for fuel, laying bare the gneiss, 
on which little or no vegetation manages to find a hold. Over other large areas the 
upper layers of peat have been removed, leaving a waste of crumbling black peat which 
does not support any vegetation. Probably all the cultivated areas have originally been 
covered with peat ; as the turbaries are exhausted and the underlying rock reached, a 


poor crop of oats and barley is grown in unfenced patches on those parts where a thin 
wash of sand and clay is present. 

Away from the villages the moors are as yet untouched, and this is the case over the 
whole of the upper part of the Bragor basin. 

Whilst the general sequence corresponds with that recorded from the Gress river 
basin, the basal beds are thicker and give more information concerning the conditions 
under which the peat began to form. 

The general sequence found over this area is as follows : — 

Characteristic Plants. 

1. Scirpus csespitosus. 

2. Eriophorum vaginatum. 

3. Betula alba. 

4. Eriophorum vaginatum. 

5. Salix Arhuscula. 

6. Empetrum nigrum. 

7. Structureless peat with seeds of Potentilla 

Comarum, Menyanthes trifoliata, Viola palus- 
tris, Potamogeton prxlongus, Phragmites com- 

Accompanying Plants. 

1. Sphagnum, Calluna vulgaris (generally scarce, but 

more plentiful near the present surface of the 

2. Calluna vulgaris. 
Rhacomitrium lanuginosum. 
PolytricJium sp., Sphagnum. 

3. Corylus Avellana. 

5. Empetrum nigrum (scarce). 

Betula nana, Potentilla Comarum, Viola paludris. 

6. Potentilla Comarum, Sphagnum. 

Sand and peaty clay. 

The fifth and sixth beds are very like the basal beds found in the Highland districts. 
Such a vegetation must have grown under sub-arctic conditions after the ice had left this 
area. It is interesting to find that the seeds of aquatic plants occur in a bed distinct 
from and below this. Beds formed of aquatic plants below the arctic vegetation 
frequently occur in other districts, such as Cape Wrath and Shetland, but they are not 
continuous like the upper layers of peat, but appear in one section and are absent in 
another only a few hundred yards away. Such beds must represent the vegetation 
growing in small marshy pools and near springs after the retreat of the ice, when 
the drier portions of the moorland were clothed with creeping willows, Empetrum, dwarf 
birch and other arctic plants. It is noteworthy that such occasional beds of aquatic 
vegetation do not appear in the hilltop and hillside peat such as the Spey-Findhorn and 
Findhorn-Nairn watersheds (3). In those localities the topography was unsuited for the 
formation of such marshy pools, and the moorland appeared to be of a distinctly drier 
type than in the low-lying districts. 

The lower forest consists of a thick bed of birch, and generally exhibits features 
which tend to show that rapid denudation was going on during its formation. The 
peat in this bed is much decomposed ; the wood is frequently badly preserved, and some- 
times little but the bark remains. Where exposed in the banks of streams, the bed is 
often seen to consist of birch bark closely pressed together without any intervening 



peat. Such features have been noted in the forest beds of other districts, and are 
discussed later in describing the geographical distribution of the strata. 

The upper forest is again absent in the basin of the Bragor river, thick beds of 
Scirpus and Eriophorum peat lying at that horizon instead of Pinus sylvestris. 

(c) Submerged Peat in Sandwick Bay, near Stornoway. — A deposit of peaty clay 
underlaid by true peat occurs under mean high-water mark in Sandwick Bay to the east 
of Stornoway. The deposit extends some distance along the beach, but it is impossible 
to say how far it extends in a seaward direction. Whilst I was in the Lews, the tides 
were not low enough to permit of sections being cut, but, owing to the kindness of Mr 
D. Mackenzie of Stornoway, I received samples of all the layers from the surface to 
the underlying rock, and an account of the sequence and thickness of the beds. 
The surface lies about 6 or 7 feet under mean high-water mark, and the total depth of the 


Fig. 1 . —Submerged peat, Sandwick Bay. Lewis. (Vertical scale £ in. = l ft.) 

deposit is 4 feet 6 inches, underlaid by conglomerate. The position of the several 
beds is illustrated in nV l. 

The peat at the base of the section is dry, hard, and shows but little structure. It 
apparently consists chiefly of the debris of birch bark and the epidermis of Phragmites 
communis rhizomes. The material is quite free from silt, and appears to have formed an 
old land surface, and the peat is certainly not the accumulation of a bog, but represents 
a forest bed, and it is clear from this that a depression of at least 1 2 feet has taken 
place since the Lower Forestian. The sand immediately overlying the peat contains 
many seeds of Zostera marina, and some traces of Sphagnum. 

4. The Mainland of Shetland. 

(Ordnance Survey— sheets 127, 128).— The Shetland Islands might be expected 
to yield results of some value owing to their separation from the mainland by a 



wide sea, which possibly has not been bridged in post-glacial times, their northern 
position, the evidence of their heavy glaciation, and the fact that there are no lofty hills 
on the mainland which would long nourish glaciers during the later stages of the glacial 

Two districts were selected for examination : — 

(«) Walls- San dness district on the west coast of the mainland. 

(b) Part of the ridge running N. and S. along the centre of the mainland north of 
Weisdale Voe. 

(a) The Walls-Sandness district is a broad peninsula of undulating moorland on 
the west coast. The highest ground is reached in Sandness Hill (817 feet) and 
Stourborough Hill (567 feet), elsewhere the moorland mostly lies between 200 and 300 
feet above sea-level. 

Most of this peninsula is formed of sandstone and flags, giving a rounded aspect 
to the hills and in striking contrast to the gneissose and quartzite rocks farther east, 
where the hills are broken in outline and covered with dry Calluna moor. The 
vegetation of the sandstones and flags where the peat is thick is composed of 
Scirpus csespitosus, Eriophorum vaginatum, stunted Calluna vulgaris, Rhacomitrium 

The peat appears to be rapidly wasting away over the whole of the mainland, and 
this is particularly marked on the hills in the Walls-Sandness region. 

Sections will first be described from the long valley running east and west 
between Sandness Hill and Stourborough Hill and drained by the Burn of Dale, 
where the peat is much channelled, the ridges between the channels being 
covered with Scirpus c&spitosus, mounds of Rhacomitrium lanuginosum, stunted 

The general sequence found over this area is as follows : — 

Characteristic Plants. 

1. Scirpus c&spitosus. 

2. Erioplwrum vaginatum. 

3. Dense light-coloured structureless peat crowded 

with the stems of Calluna. 

4. Sulix Arbusculn. 

5. Betula alba. 

6. Sphagnum and Eriophorum vaginatum. 

7. Salix reticulata, S. herbacea. 

8. Potamogp.ton pectinatus. 

Accompanying Plants. 

1. Sphagnum, Calluna (scarce). 



4. Empetrum nigrum, Betula nana, Erica Tetralix. 

5. Corylus Avellana. 

7. Betula nana. 

8. Menyanthes trifoliata, Viola palustris, Ranunculus 

repens, Equisetum sp. 

Sand and rock. 

There are certain interesting features in this sequence which merit more detailed 
description. The basal layer contains the remains of an aquatic vegetation, and although 
many of the plants have a wide range, all of them occur within the Arctic Circle at the 
present time. The relation of these aquatic plants to the overlying arctic plants is very 
similar to the aquatic vegetation underlying the arctic bed of Cape Wrath district already 



described, although the basal arctic plant beds in the two districts are not contem- 

The lowest arctic bed in Shetland contains the remains of distinctly arctic-alpine 
plants, and there can be no doubt that this bed was formed under cold conditions soon 
after the country had been glaciated. 

It is at first very difficult to say whether the basal aquatic vegetation belongs to the 
same stage as the immediately overlying arctic plants, but the fact that the aquatic 
plants were only found in a few sections renders it probable that they represent small 

rioph< im 

i na 

Sphag im 

retici ita 
jtula i na 



retic ,ta 


Peat formed from aquatic plants 
underlying the arctic bed. 

Fig. 2. — Diagram showing the general sequence of strata in the Shetland Islands. 

marshy pools which were scattered over the tundra, and therefore actually contem- 
poraneous with the lower layers of the creeping willow elsewhere in the districts. It is 
true that such evidence is not conclusive, as the oldest layers of the peat are frequently 
missing in one part of a district and present in another part of the same district, owing 
to the peat having begun to grow earlier in one place than in another. Still, there is 
one feature which makes it extremely probable that the basal aquatic vegetation is 
really contemporaneous with the arctic bed, and that is the fact that Salix reticulata 
leaves (not the stems) are occasionally present in the aquatic zone. Very frequently 
the Salix reticulata bed rests in small hollows in the floor of the moss as shown in fig. 2, 
in some cases with aquatic plants below, in other cases without. It frequently happens 


that the peat bluffs along stream sides show a band of peat formed from aquatic plants 
lying below the first arctic bed, either marking the position of a large pool or a former 
stream channel whose margin must have been fringed with M enyanthes and Potamogeton 
whilst the surrounding moorland was covered with willow and dwarf birch. 

The arctic bed is overlaid first by 14 inches of Sphagnum and Eriophorum remains, 
and then by a thick bed of Betula alba and Corylus Avellana. The trees are large, and 
represent what must have been a forest growing over the remains of the primitive 
arctic vegetation. 

But of still greater interest is the fact that above the forest remains the peat shows 
a return to arctic conditions by the occurrence of Salix reticulata, Salix Arbuscula, 
Empetrum nigrum, and Betula nana, in a perfectly distinct bed lying over the birch 
forest. This is the same sequence as I found in the Southern Uplands of Scotland, where 
an arctic plant bed overlaid the birch forest. There can, I think, be no doubt that the 
intercalated arctic bed above the birch forest in Shetland represents the same epoch as 
the intercalated arctic bed above birch forest in the Southern Uplands. 

A distinct change of climatic conditions is necessary to produce a transition from 
moorland and forest vegetation to moorland covered with creeping willow and other 
arctic plants ; and further, such climatic conditions must have operated over large areas. 
If that be so, the arctic zones in the peat are true horizons, and therefore a trustworthy 
guide to the comparative age of the beds lying below and above them. 

The sections here contain two arctic beds ; the Merrick-Kells district and Tweeds- 
muir contain only one. But the intercalated arctic bed above the birch forest is present 
in all, and I have little doubt that extended search would yield the basal arctic bed in 
the Southern Uplands. Indeed, it is quite certain that the absence of the basal arctic 
plants from districts in the south of Scotland is due to the imperfection of the record in 
the places examined, for we may be sure that a forest of birch, hazel, alder, with 
accompanying temperate plants, did not immediately follow the retreating ice. 

As the birch forest in Shetland occurs below the intercalated arctic plants, it clearly 
belongs to the lower forest period, and is contemporaneous with the birch of the districts 
in the south of Scotland. 

The lower forest bed is the only evidence of woodland conditions met with in the 
peat of Shetland, for, instead of the intercalated arctic plant bed being overlaid with the 
remains of the upper forest as in the south of Scotland, it is covered with a thick bed of 
dense structureless peat crowded with the stems of Calluna. This stratum bears every 
sign of having been accumulated under conditions which were not favourable for the 
rapid growth of peat. The Calluna stems are much shrunken, and frequently the 
interior of the stems has disappeared, only leaving the epidermis. Further, the peat 
itself is similar to the black crumbly peat now met with on the sides of old peat-hags, 
save that pressure and age have consolidated it into a hard mass. As this layer is 
traced along the banks of streams it is found to be of unequal thickness ; carefully 
examined, it closely resembles the appearance of buried peat-hags. From an examina- 



tion of many miles of this bed in the Sandness- Walls region I have no hesitation in 
saying that it marks a period of denudation during which the surface of the bed 
became wasted and channelled into peat-hags like those met with on the present surface 
of the peat. 

After this, the plant remains show that the peat has formed uninterruptedly under 
moorland or peat bog conditions down to the present time. Some of the sections taken 
in the Burn of Dale showed minor variations from the sequence just described. In 
places the birch forest rests on the glacial deposits without any arctic plants below, 
thus giving the same general sequence as the Southern Uplands. It is interesting to 
note that in such cases the birch is larger than where it is underlaid by peat, conditions 
having been more favourable for growth on the glacial clays and sands than on the peat. 

Other sections show that the intercalated arctic plants are not always present, as, in 
one or two sections, Eriophorum and Empetrum occur at this horizon instead of dwarf 

Two sections taken at the bottom of the dale and close by the burn showed the 
following sequence : — 

Characteristic Plants. 

1. Scirpus caexpitosus. 

2. Ccdluna vulgaris. 

3. Diatomaceous earth. 

4. Betula nana. 

5. Betula alba. 

6. Diatomaceous earth. 

7. Salix herbacea. 

S. herbacea x Lapponum (?) 
S. reticulata. 

8. Potamogeton pectinatus. 

Accompanying Plants. 


2. Eriophorum vaginatum. 


4. Empetrum nigrum. 



7. Betula nana. 

8. Carex ampullacea, Menyanthes trifoliata, Ranun- 
culus repens, Equisetum sp., Viola palustris. 

Whilst the three horizons are still well represented here (beds 7, 5, 4), a band of 
diatomaceous earth occurs above each arctic bed. 

During the formation of this earth the place must have been occupied by a small 
lake, and it is interesting to notice that when this was the case, peat-bog plants were 
dominant elsewhere — as beds of Eriophorum and Sphagnum occur at this level in other 

A number of sections were taken round the north and south sides of Stourborough 
Hill and Sandness Hill. In these sections the intercalated arctic bed was even better 
developed than in the floor of the Burn of Dale, but the basal arctic bed was generally 
absent, and the lower forest rested directly upon glacial deposits on the Old Red 

Sections taken farther west on Simli Field and Blonk Field showed that the lower 
forest stretched almost up to the edge of the cliffs, 200 or 300 feet above sea-level. 
Such a discovery is certainly surprising, as trees will not grow even in the most 
sheltered positions at the present time. In fact, it is even more surprising to find the 


ancient birch forest in this position than to find the remains of the upper forest far 
above modern tree limits at 3000 feet on the Grampian Mountains. 

During the examination of this region several sections were taken on the banks of 
Sma Lochs, S.E. of Stourborough Hill — a series of lochs of very irregular outline 
situated at 200 feet above sea-level. The water apparently occupies a hollow in the 
peat, as far as can be seen from the banks. Sections were made along the north side of 
the lochs. Here the banks are formed of peat which at the margin reaches a depth of 
10 feet below water-level. 

The following succession of strata was observed : — 

Characteristic Plants. 

1. Eriophorum vaginatum. 

2. CaLluna vulgaris. 

3. Betula nana. 

4. Betula alba. 

5. Structureless peat with clay. 

6. Salix Arbuscula. 

Accompanying Plants. 

1. Sphagnum, Seirpus cxspitosus. 

2. Erica Teiralix. 

6. Empetrum nigrum. 

Stone pavement. 

Whilst the general succession of vegetation is similar to that already described 
it is remarkable to find the greater depth of peat below water-level. The 
sequence and relation of the beds to the water-level in the loch is shown in 
fig. 3. In the absence of a boat upon the loch it was impossible to take borings 
to see how far the peat extends under the loch, but the horizontal strata of dwarf 
willow, peaty clay, birch forest, and arctic birch which lie below water-level, and 
stretch at least some distance under the loch, must once have been exposed. A case 
like this shows a complete reversal of peat formation, and it is difficult to see how 
such a comparatively large loch could be formed in this way. It might be pointed out 
that the water-level was rather below than above the average height, as the loch was 
examined at the beginning of September after a dry summer. There is no outlet to 
these lochs, so that the hollow in the peat cannot be accounted for by the scour of a 
large stream. Such basins in the peat, occupied by small lochs, are by no means 
uncommon ; I have observed many examples in the hilly districts of Scotland and on 
the Pennine Chain in England, but the present example is of interest on account of the 
peat upon which the loch rests, containing, first an arctic flora and then a buried 
forest, thus showing that extensive and prolonged changes have occurred during the 
peat growth. 

(b) The Ridge in the Centre of the Mainland to the North of Weisdale Voe. — From 
Weisdale Voe to Dales Voe a band of limestone traverses the gneissose rocks of the 
centre of the mainland for a distance of 12 miles N. and S., and gives rise to a range of 
smooth rounded hills with their longer axes all pointing N. and S. None of the hills 
reach 1000 feet, most of them ranging from 500 feet to 921 feet on Scalla Field. 

Sections were taken on Weisdale Hill, North Mid Field, Scalla Field, Gunti Field, 
ami Fielandringa. All these hills have a thick mantle of peat which has been furrowed 



into peat-hags ; indeed, so extensive and deep are the channels that on a clear day they 
can be well seen from a distance of 12 miles. A nearer view shows that the outline 
of the hills is ragged or like the edge of a saw, due to the depth of the numberless 

The vegetation, despite the low elevation of these hills, is unlike that covering the 
ground at 100 or 200 feet in the San dness- Walls district. Rhacomitrium lanuginosum 
is the dominant plant in places, and many of the hillsides are covered with Rhacomitrium 
moor with stunted plants of Calluna vulgaris, Erica Tetralix, Scirpus csespitosus, 
and Empetrum nigrum. 

Betula nana 

Betula alba 

peat and clay 

Salix Arbuscula 


CM <.,. 



Level of Loch 


Fig. 3. — Section on the banks of Sma Loch, Shetland. 

The general sequence over these hills is as follows : — 

Characteristic Plants. 

1. Scirpus cxspitosus. 

2. Calluna vulgaris. 

3. Salix sp., Arctostaphylos alpina. 

4. Calluna vulgaris. 

5. Betula alba. 

6. Empetrum nigrum. 

7. Salix reticulata. 

8. Potentilla Comarum. 

9. Carex sps. 

Accompanying Plants. 

1. Rhacomitrium lanuginosum. 

Kriophorum vaginatum. 

3. Empetrum nigrum (abundant). 

4. Eriophorum vaginatum, Sphagnum. 

5. Corylus Avellana. 

6. Carices. 

8. Salix Arbuscula (1) 

9. Hippuris vulgaris. 
Sand and clay. 

Comparing this sequence with that in the Sandness-Walls area it will be seen that 
whilst some of the strata contain a slightly different flora the general facies of the plant 


beds remain the same. The most important point is the presence, in this area also, of 
the lower forest and the intercalated arctic bed above. The presence of these arctic 
plants above the birch forest affords a striking confirmation of the observations I made 
in the south of Scotland, for the fact that these beds occur in the Shetland Islands tells 
a still more emphatic story of widespread changes, owing to the isolation of the islands 
from the mainland and their now unfavourable climatic conditions for forest growth. 
The presence of a buried forest on the west Shetland coast below an arctic bed shows not 
only that arctic conditions supervened after more genial forest conditions, but that the 
Atlantic cyclones must have pursued an entirely different path when that forest spread 
over the country, for the most favourable conditions of soil and temperature would not 
permit forest under present climatic conditions, and indeed a more unfavourable position 
for tree growth could hardly be found at the present time in North- West Europe. 

Previous Work in Britain on Pleistocene Plant-bearing Deposits. 

During the last thirty years much work has been done by various investigators upon 
plant remains in Pleistocene beds, and such work is admirably summarised in The 
Origin of the British Flora, by Clement Eeid. A detailed account would be out of 
place here, but reference may be made to some of the districts yielding arctic plants. 

It is obvious that the number of arctic horizons in Pleistocene and post-Pleistocene 
deposits cannot be indefinite, and therefore some of the arctic beds in peat and alluvial 
deposits may be contemporaneous with either one of the arctic zones in the Scottish 
peat mosses. For the purpose of this paper the post-glacial deposits containing arctic 
plants group themselves into two sets : — 

(a) With arctic plants lying between an upper and lower temperate bed. 

(b) With arctic plants at the base overlaid by temperate deposits. 
Type («) may be considered first. 

At Hoxne in Suffolk four distinct superposed strata containing plant remains have 
been described by a Committee appointed by the British Association (6). 
The sequence of the beds is as follows : — 

1 . Brick earth and gravel with temperate plants. 

2. Black earth with arctic plants. 

3. Lignite with temperate plants. 

4. Clay with temperate plants. 

Since the deposits fill a channel eroded through the chalky boulder clay the inter- 
calated arctic plants must belong to late glacial times long after the general glaciation 
of the country. 

It is sufficient to point out that the basal temperate beds contain such southern forms 
as Lycopus europaeus, Rhamnus Frangula, Sparganium ramosum, Sambucus nigra, 
Rosa canina, whilst the arctic bed above contains Salix polaris, S. herbacea, S. 
Myrsinites, Betula nana, to show that there is definite evidence in the south of 


England of a change from temperate to arctic conditions during the later stages of the 
glacial period. 

Fossiliferous beds of sand and peat yielding some evidence of changes in condition 
have been described from Ballaugh in the Isle of Man by a Committee of the British 
Association (7). 

In this case the evidence for a change from an arctic to a temperate flora is not so 
conclusive as at Hoxne. Sandy silt, containing Salix herbacea amongst other plants, 
is underlaid by loamy peat and Chara marl, from which arctic plants are absent, 
although most of the plants belong to species of wide distribution. The presence of 
Apus glacialis in the bed containing Salix herbacea certainly points to colder conditions 
than the present. 

Numerous sections have been described in which an arctic plant bed occurs at the 
base of the series overlaid by deposits containing a temperate flora. * 

In 1894 Bennie (8) described fossiliferous beds from Hailes, near Edinburgh, the 
epecies being determined by Clement Reid. The lower beds rest directly upon the 
boulder clay and contain Salix herbacea, S. polaris, S. reticulata, together with about 
twenty other plants, some of which certainly do not now occur in the same regions as the 
arctic willows. As the basal arctic bed occurs in a lacustrine deposit it is perhaps possible 
that it contains fossils from slightly different periods. Be that as it may, it is certain 
that the arctic willows indicate a very different type of flora to the temperate bed above 
containing such temperate forms as Sambucus nigra, Prunus spinosa, P. Padus, Rubus 
Idxus, Crataegus Oxyacantha. These plants occur mixed with others usually considered 
as weeds of cultivation, such as Chrysanthemum segetum, JEthusa Cynapium, Linum,, 
which suggests that the temperate bed belongs to a comparatively late period. If there 
is no break in continuity between the basal arctic bed and the overlying deposits 
containing temperate plants and weeds of cultivation, the evidence tends to show that 
the arctic plants belong to a period much more recent than the retreat of the ice-sheet. 

Similar arctic plant beds have been described from Corstorphine (8), Faskine (9), 
Dronachy (10), by Bennie, and from Crianlarich by Dakyns (11), the plants from the 
various sections having been determined by Clement Reid. 

Numerous other examples might be quoted from the lists given by Reid (11), but 
enough has been said to show that arctic plant beds are of wide occurrence in Britain. 
The arctic plants of Hailes, Corstorphine, Dronachy and Crianlarich apparently belong 
to the same stage, and the presence of Salix polaris in these deposits in the lowlands of 
Scotland suggests that they can hardly be younger than the first arctic bed underlying 
the lower forest in the peat mosses. 

The section at Hoxne, where an arctic bed overlies a deposit containing temperate 
plants, gives the same sequence as the south of Scotland and Shetland peat where the 
lower forest is overlaid by an arctic zone, though the fact that the Hoxne deposits rest 
upon the chalky boulder clay and the Southern Upland mosses upon morainic material of 
a later stage makes it probable that the Hoxne temperate bed is older than the Lower 



Forestian. If that be so, the arctic bed at Hoxne might still be contemporaneous with 
the first arctic bed of the peat mosses. 

There can be little doubt that the lower bed at Hoxne represents a warm interglacial 
stage, showing a very different set of conditions from the beds above containing Salix 
polaris and other arctic plants, and the events which would produce such a total change 
in the flora in the south of England must have been widespread. The suggestion made 
by Lampltjgh (12) that the arctic flora might have lingered on and re-invaded the area 
formerly occupied by a temperate flora without any widespread change in conditions, is 
surely due to a misconception of the present distribution of these plants. 

It is quite certain that Salix polaris, S. herbacea, S. Myrsinites, Betula nana, 
and such temperate plants as Rhamnus Frangula, Rubus Idmus, Rosa canina, Sambucus 
nigra, Corylus Avellana, Taxvs baccata, did not grow at the same time over the 
lowlands of the south of England. Whether the precipitation during the deposition of 
the arctic bed was great enough to produce glaciation on elevated ground in the north 
of England and Scotland the plant deposits do not tell us, but we may be certain that 
an arctic climate prevailed at that time. If then, from its position upon the chalky 
boulder clay, the Hoxne temperate bed is older than the Lower Forestian in the peat 
mosses, it carries the alternation of temperate and arctic phases to a still earlier stage — - 
though a stage which all will agree is later than the last ice-sheet. It is true that 
evidence of actual glaciation can only be given by glacial deposits, but changes of 
temperature must be reflected in fossiliferous deposits. 

A large amount of work has been done on the Continent since the first discovery by 
Nathorst in 1870 of the remains of an arctic flora in fresh- water clays near Alnarp in 
Sweden* (13). A great number of observations on such plant beds have been made 
since then, not only by Nathorst, but by other observers in Sweden, Norway, Denmark, 
Germany, and in Switzerland. The importance of the study of such stratified deposits 
before we can understand the climatic fluctuations of the Pleistocene and the modern 
distribution and character of the flora, has been fully recognised in those countries. In 
Canada and the United States hardly any work on the succession of the flora has been 
attempted, although those countries contain extensive peat deposits. The Bulletins of 
the U.S. Geological Survey contain many records of peat deposits which from their 
position in relation to the drifts and their geographical situation promise much interesting 
material towards the history of changes in the distribution of the flora. 

If the birch zone of Sweden is contemporaneous with the Lower Forestian of 
Scotland, as I suggested in 1906 (3), it is interesting to notice that the oldest forest 
zone in both countries consists of birch. But it is impossible to say whether the spruce 
and oak zones recorded from Sweden represent all the peat lying between the 

* I should have made it clear in Part II. that the first discovery of Arctic plant beds in Sweden was made in 
1870, and was entirely due to the investigations of Nathorst. These investigations were of course made in Sweden 
and not in Norway, as was inadvertently stated in Part II. A sketch of the investigations made up to 1891 is given 
by Nathorst in the paper referred to above (13). 


Lower and Upper Forestian, or the lower zone of trees in the Upper Forestian of 
the Highlands. 

In any case, the first arctic bed of Shetland and the Outer Hebrides must belong 
to the same horizon as the arctic bed first described by Nathorst from the base of the 
Swedish peat mosses. 

The Geographical Distribution of the Zones in the Peat. 

As about twenty-four districts have been investigated during the last three years 
between the south of Scotland and the Shetland Islands, Aberdeenshire and the Outer 
Hebrides, the geographical distribution of the principal zones and the variations shown by 
the same horizons in different districts may now be briefly indicated. 

The following stages are represented in the peat of the south of Scotland, the High- 
lands and the Hebrides : — 

1. Recent peat. 

2. Forest. 

3. Peat bog plants with arctic plants. 

4. Forest. 

5. Peat bog plants. 

6. Arctic plant bed. 

7. Peat bog plants. 

8. Forest. 

9. Arctic plant bed. 

It is obvious that some terminology must be used in discussing the distribution of 
these beds, and it will be simpler to provisionally name them for the purpose of descrip- 
tion and then point out the possible correlation of these strata with stages in the glacial 
period for which a terminology already exists. 

The following names will then be adopted : — 

1. Recent peat. 

2. Forest bed. \ 

3. Peat bog plants with Arctic plants. I Upper Forestian. 

4. Forest bed. ) 

5. Peat bog plants. Upper Peat Bog. 

6. Arctic plant bed. Second Arctic Bed. 

7. Peat bog plants. Lower Peat Bog. 

8. Forest bed. Lower Forestian. 

9. Arctic plant bed. First Arctic Bed. 

Strata 2, 3, 4 are all classified as "Upper Forestian," for reasons which will be 
given later. 

The First Arctic Bed. — This bed has, so far, only been met with in the Outer 
Hebrides and the Shetland Islands. Occurring at the base of all the beds, and resting 


upon the glacial deposits, it represents the arctic flora covering the peat area before the 
growth of the overlying temperate Lower Forest. The flora is best preserved in 
Shetland, where the remains of Salix reticulata, S. herbacea x Lapponumif), Betula 
nana, form a layer about 1 foot in thickness. 

Such creeping Salices occur about 2000 feet on the summits of many of the 
Highland mountains at the present time, but the conditions in Shetland must then have 
been entirely different to those which now obtain on our mountain summits ; for below 
the bed of creeping willow occur many aquatic plants such as Carex ampullacea, 
Menyanthes trifoliata, Ranunculus repens, Equisetum sps., Viola palustris, Potamo- 
geton pectinatus. These plants do not occur on the dry wind-swept summits where 
arctic willows now grow in Britain ; the flora rather suggests an arctic march or tundra 
with many scattered pools containing an aquatic vegetation. 

This bed so frequently underlying the arctic willow can hardly belong to an earlier 
stage in the peat history, as the leaves of Salix reticulata are frequently present 
amongst the aquatic and marsh plants, and have evidently been blown in from the 
surrounding moorland. The aquatic zone contains much silt, clay, and sand, whilst the 
bed of creeping willow is entirely free from such deposits. 

After the ice retreated, the ground must have been intersected by numerous rills, 
streams, and pools, in which grew Equisetum, Potamogeton and Menyanthes. Muddy 
streams frequently covered this vegetation with fine silt and sand. The drier ground 
was covered with a close growth of Salix reticulata and other arctic-alpine plants, 
which, as the pools gradually became silted up, spread over these spots also. Tt is quite 
evident that the climate of that period allowed a fairly lengthy time for the flowering 
and ripening of the seeds of such aquatic plants. In fact, the flora indicates a wet cold 
climate rather than dry conditions with an arctic temperature. Either the First Arctic 
Bed began to form some time after the disappearance of glacial conditions, or the rise 
of temperature must have been very rapid during the wane of the glaciers. The rich- 
ness of the flora indicates that the glacial stage, represented by this bed, was much later 
than the main ice sheet. 

In a recent account of the development of the Scandinavian flora, Gunnar Andersson 
(14) records the occurrence of Potamogeton filiformis, P. prselongus, Menyanthes 
trifoliata, Batrachium confervoides, and other aquatic plants in the Dryas zone, and 
infers that the Scandinavian climate of that time was not arctic like that, for instance, 
of North Greenland, but rather resembled South Greenland at the present time. This 
conclusion is in complete agreement with the features I have found in the N.W. High- 
lands. The arctic stages appear to mark periods of great precipitation, and these 
apparently alternated with comparatively dry temperate stages. 

The character of the First Arctic Bed is different in the Outer Hebrides — the only 
other district in which the peat is old enough to go back to this date. In Lewis, the 
oldest layer contains Salix Arbuscula, Betula nana, Potentilla Comarum, Empetrum 
nigrum, Menyanthes trifoliata, Potamogeton sps. 


Some of these are plants of wide distribution from temperate to arctic regions ; 
others, like Betula nana and Salix Arbuscula, have an entirely arctic or sub-arctic 
distribution and do not now occur near sea-level in Britain ; but these basal beds 
suggest less rigorous conditions than those in the Shetlands, for bog plants of wide 
distribution are the most abundant ; whilst in the Shetlands, Salix reticulata is the 
characteristic plant. 

The plants of the First Arctic Bed must have immigrated to Scotland, either across 
the North Sea, or through England. The incoming flora would probably pass along 
the higher ground, while the valleys and lowlands were still covered with the shrinking 
remnants of the local ice-sheets. The evidence from the mosses in Lewis certainly 
suggests that the incoming arctic flora barely established itself in the extreme west of 
Scotland before the Lower Forest overspread the peat bogs, as the typical arctic plants 
represented in Shetland are either absent or poorly represented in the corresponding 
zone of the Hebrides. 

The Lower Forestian. — Considerable interest attaches to the flora of this zone, owing 
to the absence of any arctic or sub-arctic plants, suggesting that a complete change 
of conditions took place between the first and second arctic beds during which the 
arctic flora withdrew from the peat areas. 

In the Southern Uplands, Hebrides and the Shetlands, the Lower Forest is well 
preserved, the trees being Betula alba, Corylus Avellana, Alnus glutinosa. The list 
of accompanying plants is not large, as the peat is chiefly formed of wood debris and 
bark. Inferences as to temperature drawn from these trees alone are inconclusive, but 
certain features presented by the distribution of the forest-bed bear directly upon the 
general climatic conditions of that time. The presence of a buried forest extending to 
the western coast of Shetland points indubitably to the fact that the path of the Atlantic 
cyclonic systems must then have been different to that of the present day. Trees are 
now entirely absent from Shetland, although many attempts have been made to form 
plantations in sheltered spots. Planted in gardens, ash, rowan and birch grow as high 
as the surrounding walls, and are then stunted by the winds, so that the trees appear 
to be shorn off at the level of the protecting wall. If this takes place at the inland 
and more sheltered spots, it is still more difficult to account for the presence of the 
buried forest near the cliffs of the west coast, where the full force of the Atlantic gales 
is felt, as they succeed one another during the greater part of the year. As the mean 
annual temperature of the Shetlands is perhaps higher than many parts of Britain 
which are tree-clad, the absence of trees in these islands must be due, not to tempera- 
ture, but to the force of the salt-laden winds. 

The accompanying flora of the Lower Forest contains many plants of wide distribu- 
tion which do not give any definite evidence with regard to temperature. The zone 
also contains other plants, which are not usually found accompanying arctic plant 
associations, such as Ajuga reptans, Ranunculus lingua, Lychnis diurna, Salix 


But the most significant fact is that none of the characteristic arctic plants of the 
First and Second Arctic beds are to be found in the Lower Forest Bed, though quantities 
of material from the south of Scotland, the Hebrides and the Shetlands, have been worked 
through. If it still be urged that there is no direct proof that the Lower Forest Bed was 
accumulated under temperate conditions, one may ask what was the reason for the 
disappearance of the characteristic plants of the First Arctic Bed during this period, and 
why should these typical arctic species again appear as the strata above the Lower 
Forest are reached ? 

The upper limit of tree growth has been used both in Britain and on the Continent 
as the approximate boundary between the sub-alpine and alpine floras, and it is found 
that a considerable difference in altitude separates the zone of natural woodland from 
the alpine associations. 

From this, the First Arctic Bed appears to be the result of the depression of the 
present alpine-arctic boundary in Britain to the extent of about 2000 feet, whilst the 
Lower Forest represents the return of that boundary to at least its present altitude. 

The chief features in the distribution of the First Arctic Bed and the Lower Forestian 
are shown in Plate II. The absence of these two beds from the Highland districts is 
remarkable. Either conditions were not favourable for peat formation in those areas 
during the First Arctic, the Lower Forestian, and the Lower Peat Bog period, or peat 
formed during those periods was subsequently removed either by glaciation or by 
atmospheric agencies. Further areas must be examined before it is possible to say 
which view is correct. 

At the same time it must not be forgotten that water action frequently plays a great 
part in the removal of peat beds. Many cases have come under my notice in various 
parts of Scotland and the north of England where the older beds along a stream-side 
have been swept away and replaced by mounds of sand and gravel. This is illustrated 
in fig. 4, showing a sharp bend in the Abhuinn-a-Coire-Bhuig in Easter Eoss, where all 
the older peat beds have been washed away and replaced with sand and gravel. This 
event occurred before the Upper Forestian, and most probably during the Upper Peat 
Bog stage. It is not, of course, suggested that the absence of the older peat beds in the 
Highlauds can be due to water action of this character, but the example is given in 
order to emphasise the many occurrences which may be responsible for the absence of 
the older beds of peat in some districts. It is quite evident that peat formation is 
dependent upon many factors, such as temperature, humidity of climate, drainage, 
character of soil and the flora, or the peat deposits of Britain would not be so local in 

No trace of Pinus sylvestris has been found in the Lower Forest, although it is the 
most characteristic tree of the Upper Forest. Tf native in Scotland at that time it is 
difficult to see why it should not occasionally have spread over the peat-covered areas, 
but neither in the south of Scotland nor in the Hebrides and Shetlands is any trace 
of it to be found. 



The Lower Peat Bog— This zone is present in the low-lying Wigtownshire mosses 
and the other areas examined in the Southern Uplands, in Skye, the Outer Hebrides, 
and in the Shetlands. In the Southern Upland areas and the Shetlands it exhibits the 
greatest uniformity, the bulk of the peat being formed from the remains of Sphagnum, 
Scirpus csespitosus, Eriophorum vaginatum, E. angustifolium, Molinia cserulea, Carices. 
Calluna is scarce or absent, and the peat areas at this time seem to have formed extremely 
wet Scirpus and Sphagnum moors, and. trees appear to have been entirely absent. The 
peat bears every sign of having been formed under conditions eminently favourable for 
peat growth, and although this zone frequently attains a thickness of 5 to 7 feet it is 

1 ss 





Pinus Eriophorum 

sylvestris and 



Empetrum Salix 

Fig. 4. — Showing the replacement of the older beds of peat by fresh-water sands and gravels at sharp bends along stream-sides. 

1400 feet on the Abhuinn-a-Coire-Bhuig, Easter Ross. 

probable that it was formed in less time than the First Arctic Bed, which is only 18 
inches in average thickness. 

In the low-lying districts of Wigtownshire, Skye, and the Uist a different type of flora 
overspread the peat during this stage, for the Lower Peat Bog there consists of the re- 
mains of Phragmites communis, Equisetum sps., Menyanthes trifoliata. In fact, these 
areas must have been covered with a series of shallow lakes and swamps, possibly explained 
by the fact that the peat rests upon stiff clays, whilst many of the Southern Upland 
mosses rest upon sand and gravels. In Lewis, the Lower Peat Bog approximates more 
to the Southern Upland type, being composed of Sphagnum, Eriophorum angustifolium, 
Erica Tetralix. 

A noticeable feature is the fact that the Lower Peat Bog tends to thin out as it is 
traced across Scotland from the south to the north. This is a phenomenon of some 
interest, particularly when considered in connection with the significance of the succession 


of the peat strata. It is evident that these successive beds represent changes in plant 
distribution on a great scale, or occupying presumably considerable spaces of time. The 
Lower Peat Bog stage undoubtedly represents a great invasion of North Britain by 
swamp and bog plants which previously had their centre either in the south or east of 
Britain. If that be so, the arrival of this type of flora would be much later in the extreme 
N. of Scotland than, for instance, in the Southern Uplands. It is perhaps not without 
significance that in Shetland the Lower Peat Bog thins out to a bed only a few inches 
in thickness or a layer of diatomaceous earth. The centre of distribution of these 
swamp and bog associations could hardly have been in the north of Britain during the 
growth of the lower forest, and the farther the areas were from this centre the longer it 
would be before they were colonised. 

The Second Arctic Bed. — The Lower Peat Bog is succeeded by a flora composed 
mainly of typical arctic plants and, as will be seen from Plate II. , is represented in 
every district so far examined along the line A B on Plate I. from the Southern Uplands 
to the Shetlands. In the south of Scotland the zone contains Salix reticulata, S. 
herbacea, Loiseleuria procumbens, Empetrum nigrum, Arctostaphylos alpina ; but in 
the Highlands and in the Shetlands Salix reticulata, Betula nana, Empetrum nigrum, 
Salix Arbuscula, Erica Tetralix. In these areas the Second Arctic Bed occurs between 
the Lower and the Upper Peat Bog, but in the areas which lie within the Highlands no 
older beds lie below this zone, and it rests directly upon sands, gravels, and clays. In 
such districts the bed is thicker and contains a richer flora than in the Shetlands and 
south of Scotland. 

In the areas stretching from the Grampian mountains to Cape Wrath the peat of 
this bed contains Dryas octopetala, Salix reticulata, S. herbacea, Arctostaphylos 
alpina, Betula nana, Veronica alpina, Lychnis alpina, Salix Arbuscula, Carex sps., 
Equisetum sps., Empetrum nigrum, Potentilla Comarum, Viola palustris, Menyanthes 

In the district near Cape Wrath these plants are frequently underlaid by a thin 
seam of peat a few inches in thickness, composed of the remains of aquatic or marsh 
vegetation, consisting of Ranunculus repens, R. Jlammula, Potamogeton natans, 
P. prsdongus, P. rufescens, Menyanthes trifoliata. This seam is not continuous but 
appears to underlie the Salix beds in small patches and frequently in depressions of 
the ground, and marks the position of small swampy patches during the Second 
Arctic Period. 

In the districts of the extreme west, along the line B C, Plate I., such as Skye, N. 
Uist and Lewis, the Second Arctic Bed is not recognisable, and the First Arctic Bed is but 
poorly represented. After the close of the Lower Forestian the general type of vegeta- 
tion over these peat mosses seems to have remained very uniform, through the Lower 
Peat Bog, Second Arctic Stage, and the Upper Forestian to the present day. During 
those periods the mosses in the Hebrides must have resembled the tundra of Northern 
Siberia described by Sewell (15). 


The observations from the Shetlands illustrate the same feature, for although the 
Second Arctic Bed is present, yet the Upper Forest does not seem to have spread over 
the peat areas. Shortly stated, the farther one passes to the edge of the Continental 
plateau the more uniform does the flora appear to have remained since the later stages 
of the glacial period. This may imply either that the climatic conditions have been 
more uniform in those regions, or that the successive waves of vegetation passing over 
the mainland did not penetrate there on account of natural barriers to immigration. 
The peat does not afford any direct evidence upon these points, for whilst the Second 
Arctic Bed was being formed on the mainland, the peat mosses of the Hebrides were 
covered with marsh consisting of Phragmites communis and Equisetum, which might 
well occur under cold or under temperate conditions. 

The Upper Peat Bog is developed in all districts, and the flora is generally very 
similar to that of the Lower Peat Bog. In some of the Highland areas, such as the 
Spey-Findhorn watershed, Findhorn-Nairn watershed, Coire Bog, Cape Wrath district, 
Rannoch Muir, remains of small Betula alba shrubs occur in the lower parts of this zone. 
The wood is very local in occurrence, whilst the main mass of the peat is formed 
of Scirpus csespitosus, Eriophorum angustifolium, E. vaginatum, and Sphagnum. 
Evidently these small clumps of shrubby birch were soon displaced, as no wood is met 
with above the first foot or so of the Upper Peat Bog. In the Hebrides and the 
lowland mosses of Wigtownshire the Upper Peat Bog is simply a continuation of the 
Lower Peat Bog, there being no layer of arctic plants between. 

The Upper Forest Bed. — This zone is confined to the mainland of Scotland from 
the lowland mosses of Wigtownshire to the valley of the Dionard, south of the Kyle 
of Durness. Whilst the flora exhibits remarkable uniformity, some interesting and 
important modifications appear as the bed is traced northwards through Scotland. In 
the areas of the south of Scotland the forest consists of Pinus sylvestris, with the 
exception of some parts of Tweedsmuir, where Betula alba replaces pine. As soon 
as the Highland areas are entered the Upper Forest divides into two distinct zones 
separated by 1-3 feet of peat in which no wood is found. In the Spey-Findhorn, 
Findhorn-Nairn, Coire Bog districts, the peat between the upper and lower layers of 
this forest is formed from Sphagnum, but in areas examined in the Grampians and in 
Assynt, Betula nana and Salix Arbuscula are abundant between the two layers of 
forest remains. The presence of these plants in some districts, and of Sphagnum beds 
in others, between two layers of well-developed pine forest, is significant, and points to 
a decided break in forest conditions at this time. This phenomenon is too widespread 
to be due to local causes, for it is a constant feature in all the Highland areas 

Betula alba replaces Pinus sylvestris in the lower zone of the Upper Forest on the 

Grampians, in Assynt, and over much of the Caithness-Sutherland border ; in other 

words, at high elevations and in the extreme north. 

It is remarkable that until the Upper Forest is reached no extensive beds of Calluna 


are found. The Vaccinium-Calluna association, now so constant a feature on moorlands 
in Scotland and England, appears to be a comparatively recent feature. 

The Peat above the Upper Forest. — The peat of varying thickness lying above the 
Upper Forest presents the same features as the Lower and Upper Peat Bog. Occasional 
beds of Calluna or Eriophorum make their appearance, but are not continuous over 
large areas, and are evidently due to local causes. The general flora is of a distinctly 
wetter type than that characteristic of many of the peat areas of the present time, as it 
consists mainly of Scirpus csespitosus and Sphagnum. Although this zone does not 
indicate any widespread changes in the distribution of the flora it shows some features of 
oecological interest in connection with the present denudation of the peat ; these will be 
dealt with in another paper. 

The Relation of the Strata to the Glacial Succession. 

It is evident that the peat mosses do not give any information about the chain of 
events during the maximum glaciation of the country, and it is immaterial whether we 
regard the morainic material upon which so many of the older peat mosses in the south 
of Scotland rest, as the deposits of the waning ice-sheet of maximum glaciation, or as 
the deposits of an entirely distinct glacial stage, separated from the ice-sheet by a warm 
inter-glacial phase. The fact remains that the First Arctic Bed contains an arctic-alpine 
flora which existed over wide areas near sea-level. It is of course impossible to say 
from the evidence of the plants whether all traces of glaciation had vanished from 
Britain at that time, or whether certain regions were still under ice. An entirely 
different flora makes its appearance in the Lower Forestian ; not only are all arctic- 
alpine plants absent, but the flora is made up of well-grown trees and an assemblage 
of plants quite typical of any marshy lowland forest region in Britain at the 
present time. 

If it be contended that the Lower Forestian is due to local changes in climate — such 
as variation in snowfall, direction of wind — how is it that the bed maintains its 
character, not only in the south of Scotland, but also in the Outer Hebrides and Shetland 
Islands ? 

If the succession of the First Arctic by the Lower Forestian, Lower Peat Bog and 
Second Arctic Bed were really due to local causes, widely separated districts should show 
different successions ; while, for instance, tundra vegetation would be represented in one 
district the remains of a forest vegetation would occupy the same horizon in another. 
But this is not the case, for nothing is more striking than the continuity of the horizons. 
The chief points for consideration may be stated as follows ; — 

(a) First Arctic Bed. Lower limit of arctic-alpine vegetation depressed nearly 
to sea-level. 

(b) Lower Forestian. Upper limit of deciduous trees raised to at least its present 


(c) Second Arctic Bed. Lower limit of arctic-alpine vegetation depressed nearly 
to sea-level. 

(d) Upper Forestian. Upper limit of trees raised to over 3000 feet above 

These changes are post-glacial in the sense that they occurred later than the last 
ice-sheet. It is difficult to reconcile these changes with the " gradually waning 
glaciation, during which there were occasional local advances of the mountain-glaciers 
in their glens due to temporary increase of snowfall" suggested by Lampltjgh (12). 

The beds rather indicate a definite sequence of events which took place simultane- 
ously over the whole of Britain. Local changes of snowfall, and local retreat and 
advance of glaciers, almost certainly occurred during the later stages of the glacial 
period, but these could hardly bring about the widespread alternate depression and 
elevation of the limits of an arctic-alpine flora shown by the peat. 

It matters little whether these stages are spoken of as local changes during the 
waning glaciation or are expressed as glacial and inter-glacial stages ; the fact remains 
that the climatic fluctuations were lengthy and pronounced enough to change the 
distribution of the flora in the north of Britain. 

As the arctic beds contain plants indicative of cold and wet conditions, the 
probability is that glaciation took place in elevated regions during their deposition. 
The Lower Forestian contains a flora indicating conditions at least as temperate as the 
present day, and the Upper Forestian shows an elevation of the upper limit of forest 
far above that of the present day. 

In Sweden a similar elevation of the upper limit of forest to the extent of 1100 
feet (300 m.) took place during post-glacial times, and Gunnar Andersson concludes 
that the arctic-alpine flora then disappeared from the lower mountain ranges. My 
own observations in Scotland would place the difference between the upper limit of 
trees during the Upper Forestian and at the present day at not much less than 2000 
feet in the Highlands ; in Cumberland and Westmorland at rather more. It is 
possible that glaciers still lingered on in many parts of the Highlands during the earlier 
stage of the forest periods ; this would be more likely in the case of the Lower Forestian 
than the Upper Forestian. 

The question of climate during the Pleistocene epoch has been discussed by Harmer 
(16) in an interesting and suggestive paper, and the conclusion is reached that a change 
in the direction of prevalent winds, due to different relative positions of areas of high 
and low barometric pressure, may account for the comparatively low temperature and 
high precipitation during some stages of the Pleistocene period. The observations from 
Shetland certainly show a difference in the direction of the prevalent winds during the 
[Lower Forestian. 

While it is difficult to reconcile the several stages in the peat with the theory of 
a single glaciation, the whole of the peat beds agree very closely with the scheme of 
classification proposed by Geikie (17). In that scheme the First Arctic Bed and Lower 



Peat Bog would mark the gradual passage of the Mecklenburgian stage into the Lower 
Forestian, the Second Arctic Bed would represent the Lower Turbarian, and the Upper 
Forest in the peat would correspond with the Upper Forestian or Fifth inter-glacial 


All the Scottish peat mosses show a definite succession of plant remains. The 
oldest, in the south of Scotland and the Shetland Islands, have an arctic plant bed at 
the base. This is succeeded by a forest of birch, hazel, and alder containing temperate 
plants. A second arctic plant bed occurs above the Lower Forest and is overlaid in all 

5ea /ere/ 

Lower limit of 
I" Arctic be J. 

Upper limit of 
lower forest 

Lower limit of 
Z°* Arctic bed 

Upper limit of 
Upper foresf 

Fig. 5.— Broken lines represent the present limits of Arctic-Alpine and forest vegetation ; thick 
lines show the limits attained by these formations during the growth of the peat. 

districts (except the Hebrides, Cape Wrath, and the Shetland Islands) by an Upper 
Forest covered by several feet of peat-bog plants. 

The altitudinal limits of these successive zones is shown in fig. 5. It will be seen 
that the two arctic beds descend to within about 150 feet of sea-level, whilst the Lower 
Forest rises to 1500 feet and the Upper Forest to close upon 3500 feet. The present 
lower limit of an arctic-alpine vegetation in Britain is taken as 2000 feet, and the upper 
limit of forest (pine and birch) as 2000 feet, although in many districts it lies consider- 
ably lower than this. The upper limit of the Lower Forest is given at 1500 feet, owing 
to the fact that all the beds below the second arctic zone are wanting above this 
altitude ; but as the Lower Forest is present in many exposed situations, such as the 
west coast of Shetland, it is probable that the upper limit of forest lay considerably 
above this level. 


In conclusion, I wish to express my thanks to Dr Horne, F.R.S., for much valuable 
help and advice and for the use of Geological Survey maps, both published and in 
manuscript. My best thanks are also due to Professor Geikie, LL.D., D.C. L, F.R.S., 
for many suggestions. 

The scientific expenses of this investigation have been defrayed by a grant from the 
Government Grant Committee of the Royal Society, London. 


(1) Andersson, Gunnar, " Die Geschichte der Vegetation Schwedens," Engler's Bot. Jahrbikher, Bd. 

v. Heft 3, 1896. 

(2) Pethybridge, G. H., and Praeger, R. Lloyd, "The Vegetation of the District lying South of 

Dublin," Proc. Boy. Irish Acad., vol. xxv., Sect. B, No. 6, 1905. 

(3) Lewis, F. J., "The Plant Remains in the Scottish Peat Mosses," Part II., Trans. Boy. Soc. Edin., 

vol. xlv. part ii. No. 13. 

(4) Smith, W. G., "Botanical Survey of Scotland," III. and IV., Forfar and Fife, Scot. Geographical 

Magazine, Dec. 1904. 

(5) Geikie, J., " Buried Forests and Peat Mosses of Scotland," Trans. Boy. Soc. Edin., vol. xxiv. 

(6) Reid, C, " Report of Committee on the Relation of Palaeolithic Man to the Glacial Epoch," Bep. 

Brit. Assoc, for 1896. 

(7) Reid, C, " Report of Committee on Irish Elk Remains in the Isle of Man," Bep. Brit. Assoc, for 


(8) Bennie, J., "Arctic Plants in the Old Lake Deposits of Scotland," Ann. Scot. Nat. Hist., 1894. 

(9) Bennie, J., "On the Occurrence of Peat with Arctic Plants in Boulder Clay at Faskine, near 

Airdrie," Trans. Geol. Soc. Glasgoio, vol. x., 1895. 

(10) Bennie, J., "Arctic Plant Beds in Scotland," Ann. Scot. Nat. Hist, 1896. 

(11) Reid, C, Origin, of the British Flora, 1899. 

(12) Lamplugh, G. W., Presidential Address to Geological Section, British Association, York, 1906. 

(13) Nathorst, A. G., " Ueber de gegenwartigen Standpunkt unserer Kenntnis von dem Vorkommen 

fossiler Glacialpflanzen," Bihang till K. Svenska Vetenskaps Akad. Handlingar, Stockholm, 17 
(1891), III. 

(14) Andersson, Gunnar, "Die Entwicklungsgeschichte der Skandinavischen Flora," Besultats 

scientifiques du Congres international de Botanigue, Wien, 1905. 

(15) Sewell, P., "The Flora of the Coasts of Lapland and of the Yugor Straits," Trans, and Proc. Bot. 

Soc. Edin., 1888-89. 

(16) Harmer, F. W., "The Influence of the Winds upon Climate during the Pleistocene Epoch," Geol. 

Jour., vol. lvii., 1901. 

(17) Geikie, J., The Great Ice Age, 3rd edition, 1894. 


Plate I. Map of Scotland, showing the areas investigated during 1904, 1905, 1906. 
Plate II. Sections I. and II., illustrating the sequence along the line A B, C D, respectively on Plate I. 
Plate III. Plant remains from peat deposits. 

Fig. 1. Betula nana leaves from between the two layers of upper forest, Rannoch Muir 
(a b x 8 ; c nat. size). 






2. Salix reticulata leaves from the first arctic bed, Shetland. 

3. Arctostaphylos alpina seed from the basal arctic bed, Spey-Findhorn watershed. 

4. Ranunculus repens achene from below the second arctic bed, Cape Wrath. 

5. Hippuris vulgaris seed from basal layers, Shetland. 

6. Potentilla Comarum achene from basal layers, Shetland. 

7. Empetrum nigrum seed from second arctic bed, Spey-Findhorn watershed. 

8. Potamogeton natans from below the first arctic bed, Shetland. 

9. Corylus avellana nut from the lower forest, Skye. 

10. Alisma plantago fruit from basal layers, Moorfoot Hills, a, view from above ; b, side 


11. Pinus sylvestris cone from upper forest, Coire Bog, Easter Ross. 

12. Salix Arbuscula leaf from basal layers, Coire Bog, Easter Ross. 

13. Viola palustris seeds from basal layers, Lewis. 

14. Menyanthes tri/oliata seed from the lower forest, Shetland. 

15. Ajuga reptans seed from lower forest. Tweedsmuir. 

Trans. Roy. Soc. Edin. Vol. XLVI. 

Lewis: Plant Remains in the Scottish Peat Mosses. Part III. — Plate I. 


The Peat Mosses investigated during 1904-1906 


/St A UOS 

>0 20 JO 


Trans. Re 

F. J. Lewis 

Plant Remains in Scottish Peat Mossee 


Part 111 , Plate II. 

■ItituAe aWvt 









(3ooVvCo<l»\) I I 

! ! 





Salix reticulata . Salix herbacea . Salix Arbuscula . 
Arctostaphylos alpina . Dr yas octopetala . Betula nana . 
L ychnis alpina . Veronica alpina . 
Sedum Rhodiola. Empetrum nigrum . Sa&ina nodosa . 


Betula alba . Corylus avellana . Almjs_gl_uJ^no^a. 
AjuA a reptans . Ranunculus lin g ua . Pmus sylvestris . 
Salix purpurea . Viburnum opulus ( Cross Fell district) 
Elatine hexandra (Cross Fell district) 


oy Soc.Edm r Vol.XLYI. 

F. J. Lewis: Plant Remains in Scottish Peat Mosses. — ■ Plate III. 

x 8 M& a 

Fig. 2. 

Fig. 1. 

Fig. 3. 

* 3 il # 

Fig. 4. 

Fag 5. 


Fag. 6. 


SJ x 3 

Fag 7. 

Fig. 8. 

Fig- 9. 

9 x8 

Fig. 10. 


x 8 


Fig. 12. 


Fig 13 

x 3 

x 3 

Fig 14. 


Fig. 15. 

71 c Fs.Tlane l> fci-s^iinc Lith Edm* 

<f> A & Jk *> 

Trans. Roy. Sac. Edin. 

Vol. XLVI. 

Lewis : Plant Kemains in the Scottish Peat Mosses. Part III. — Plate IV. 

Fig. 1. — The Upper Forest zone (Pinus sylvestris), Rannoch Muir. 

Flu. 2. — Rannoch Muir. The vegetation consists of Rhacomitrium laniujinusum, 
Scirpus cccspitosus, and Callitna. 


surface of the foetus it communicates directly with the amniotic vesicle. When later 
the opening of the bowel differentiates from the urinary orifice, both anus and the 
meatus urinarius are in direct communication with the vesicle. 

In reptiles, birds, and mammals another vesicle filled with fluid develops to a greater 
or less extent. This is the allantois. It is an outgrowth from the posterior end of the 
gut, which grows out in the mesoblast surrounding the amnion, and it may finally 
separate the amnion from the periphery of the ovum. From the first it is essentially 
a continuation of that part of the hind gut which will afterwards form the urinary 
bladder of the foetus. In the wall of this vesicle, in most mammals, the blood-vessels 
of the embryo form a connection with the lining membrane of the uterus and produce 
a placenta. 

In most of the lower mammals both vesicles are well developed ; but in man and the 
anthropoid apes the allantois attains a very insignificant size and in man takes no part 
in the formation of the placenta. 

The source of the fluids which fill these vesicles is still doubtful. In reptiles and 
birds they are necessarily fcetal in origin. The occurrence of urea and uric acid in the 
allantoic fluid of these animals, as was long ago demonstrated by Jacobson and Prevost 
(Gusserow, Arch. f. Gynaekologie, Bd. iii. p. 244, 1872), shows that in them the 
allantoic fluid is derived, in part at least, from the kidneys. 

In mammals the question is not so simple, since the membranes surrounding one or 
both of the vesicles are in close relationship to the mucous membrane of the maternal 
uterus. The possibility is thus suggested that one or both fluids may be derived from 
the mother — a view which has been upheld by one set of authorities, and which is very 
commonly accepted in modern text-books of obstetrics. Other observers maintain that 
the fluids are fcetal in origin in the mammal as in the sauropsida. 

A study of the anatomical relations of the membranes makes it difficult to accept 
the former hypothesis without very conclusive evidence in its favour. In most mammals 
the membrane of the allantois alone is in relation to the maternal mucous membrane ; 
but it is always definitely separate from it, and is easily detachable. 

The only connection between the fcetal and maternal circulation is through the 
placenta, which, although it varies greatly in distribution and complexity, is always 
essentially the same in structure — consisting of tufts of fcetal blood-vessels, carried out 
in the allantoic mesoblast, which have thrust their way into blood-spaces in the mucous 
membrane of the maternal uterus, so that the fcetal blood is brought into intimate 
relation with the maternal blood. 

In all mammals fcetal blood-vessels run throughout the wall of the allantois — 
both in that part lying against the mucous membrane of the uterus, but separated 
from it by the chorionic amnion, and also in that part which surrounds the 
amniotic sac. 

In man, while in the very early stage of development the ovum is attached all 


round to the maternal mucosa by primitive chorionic villi, at the end of the third 
month of pregnancy the connection between chorion and mucosa breaks down as the 
placenta forms, and in all later stages is loose and non-vascular, except at the placental 
site. The connection between the outer layer of the amnion and the inner layer of the 
chorion is also loose, and can be readily separated even at full time. 

Viewing the question from the anatomical standpoint, it is difficult to see how the 
fluids in these vesicles can be derived from the maternal blood. 

2. Pathological Evidence. 

This has been chiefly derived from the study of polyhydramnios and oligohydramnios 
in the human ovum. The most noteworthy of the recent observations on this subject 
is Silberstein's (Arch. f. Gfyn., lxviii. 607, 1903). From a study of a 5^-months 
uni-ovular twin pregnancy, in which one of the sacs contained 6^ litres of fluid, while 
the other contained very little, he comes to the conclusion that part at least of the 
human liquor amnii is derived from the foetal kidneys. In the sac containing the 
large quantity of fluid the foetus was much larger than the other — 783 grms. as 
contrasted with 488 grms. The heart of the polyhydramniotic foetus was 7 '5 grms., 
and that of the oligohydramniotic one, 3 grms. ; the kidneys, 2 '9 grms. each, as against 
1 grm. each; and the bladder, 2'1 grms., as compared with 2 grms. The glomenuli 
in the kidneys of the larger foetus had a diameter of 97 to 115 micros, while 
those of the smaller were only from 65 to 66 micros. 

He concludes that, owing to the arrangement of the vascular areas of the two 
foetuses in the common placenta, the larger foetus had an extra supply of blood, 
part of which was directly diverted from the smaller one, and that as a result all 
the organs had hypertrophied, probably beginning with the heart. He considers the 
excess of liquor amnii in the sac as due to the consequent increased activity of the 

The mere fact of only one of the twin amniotic sacs having an excess of fluid is, of 
course, strong presumptive evidence of the foetal origin of that fluid ; but the evidence 
afforded that the fluid was secreted by the foetal kidneys is not so convincing. Cases 
of hydramnios have been recorded in which there has been obstruction in the urethra 
and hydronephrosis in the foetus (Jaggard, Am. Jour. Oust., xxix. p. 432, 1894). Like 
conditions in the foetus have been found associated with oligohydramnios, as in cases 
recorded by Ballantyne (Ed. Med. Jour., xl. 858, 1895) and Blackwood (Ed. Med. 
Jour., xli. 919, 1896); so that no definite conclusion, either in favour of or against the 
renal origin of the fluid, can be drawn. 

In other cases the hydramnios has been found associated with conditions interfering 
with the circulation in the umbilical cord, such as excessive torsion, thrombosis of the 
vein, and cirrhosis of the liver (Opitz, Central, f. Gi/n., 1898, p. 553). It is well 
known that hydramnios is often associated with malformations in the foetus, such as 


spina bifida, auencephaly, etc. But which is the primary condition has never been 
definitely ascertained. Altogether, very little light has been thrown upon the origin 
of the fluids from a study of pathological conditions in the human subject. 

3. Experimental Evidence. 

(1) The Removal of the Fluids. — It seemed to us that the effects of withdrawing 
the fluids might throw light upon the question of their origin, and one of us carried out 
a series of experiments on rabbits (Watson, Jour, of Obstet. and Gyn. of the British 
Empire, Jan. 1906). The fluids were withdrawn from the uterus by an aspirator 
needle (the abdomen having been opened) at different periods of pregnancy. In the 
early stages of pregnancy, the fluid removed was yolk-sac fluid, and in the middle and 
later stages liquor amnii. In both cases the withdrawal resulted in immediate death of 
the foetus, and this was followed by a degeneration of the foetal part of the placenta. 
The maternal part of the placenta, on the other hand, continued to grow and to undergo 
the various histological changes which occur under normal circumstances. In no case, 
however, was there any re-formation of the liquor amnii. This would appear to indicate 
that in the rabbit the fluid must be foetal in origin, since its secretion is arrested by the 
death of the foetus, in spite of the fact that the foetal part of the placenta does not show 
marked signs of degeneration for a few days, and the maternal part remains apparently 
normal up to the time it is cast off. 

(2) The Passage of Chemical Substances from the Mother to the Fluids. — A large 
number of experiments have been recorded showing that chemical substances injected 
into the maternal circulation may appear in the fluids ; and from certain of these the 
conclusion has been drawn that the fluids are formed directly from the maternal blood. 
The best known of these are by Zuntz {Pflilg. Arch., Bd. xvi. p. 548, 1878). 

The method he employed was that which had been already used by Gusserow in 
his classical investigations {Arch. f. Gyn., Bd. iii. p. 241, 1872), and according to 
Gusserow by even earlier observers, e.g. Meyer (Bischoffs Entwickelungsgeschichte, 
p. 515). The substance used by Zuntz was sodium sulphindigotate, which he injected in 
large quantities into the veins of pregnant rabbits. He states that the amniotic fluid 
was coloured blue, while the kidneys and other tissues of the foetus were free of the 
pigment, and that, when he killed the foetus by injecting into it through the uterine 
wall caustic potash before injecting the pigment, in one experiment at least the pigment 
was detected in the fluid. 

These observations were confirmed by Wiener for rabbits in the later stage of 
pregnancy {A rch. f. Gyn., Bd. xvii. p. 24, 1881, and Bd. xxiii. p. 183, 1884). He further 
showed that the passage of pigment into the foetal fluids was more rapid when the 
kidneys of the mother were first removed. In two pregnant dogs he entirely failed to 
find the pigment in the foetal fluids, and he also failed in rabbits during the early stages 
of pregnancy. 


It was thus in rabbits only that the passage of this colouring-matter to the fetal 
fluids, without its presence having been detected in the foetus, was observed by these 

Krukenberg (Arch. f Gyn., Bd. xxii. p. 39, 1884) criticises their conclusions, 
pointing out that the recognition of sulphindigotate of soda in tissues is not easy, as 
was indicated by Heidenhain (Schultzes Arch., Bd. x. p. 35), who stated that, by 
injecting very small quantities, the urine may be made distinctly blue, while it is 
impossible to recognise the blue colour in the tissues. Krukenberg has repeated 
this experiment on the rabbit, and finds that, while the urine is distinctly blue, it is 
impossible to detect the pigment in the kidneys and liver. He does not prosecute 
the question of whether this is simply due to the small quantity of pigment, or 
whether "organic substances in close contact with small quantities of the pigment 
can decolorise it." 

He further criticises the experimentum crucis of killing the foetus by injecting caustic 
potash and then finding a slight blue coloration of the amniotic fluid. Certainly it 
seems possible that a traumatic communication might thus have been established 
between the maternal vessels and the fluids — a conclusion which is supported by the fact 
that in all experiments but one the red coloration of the fluid from extra vasated blood 
prevented the detection of the blue pigment. 

Opposed to the conclusion of Zuntz and Wiener are the experiments of Gusserow 
of injecting benzoic acid into the maternal blood and finding hippuric acid in the 
fcetal fluids. Accepting the conclusion of Bunge and Smiedeberg that the synthesis of 
this substance takes place in the kidney, this observation favours the view that the fluids 
are produced from the embryonic kidney. 

II. Present Investigations. 

The unsatisfactory state of our knowledge as to the source of these fluids induced us 
to take up the question, and we have studied it by two different methods — first, by a 
continuation of the method of Gusserow, Zuntz, and others ; and secondly, by a study of 
the chemistry of the fluids. 

1. The Part played by the Foetus in the Passage of Substances from the Maternal 

Circulation to the Fluids. 

The objections to the use of the blue sodium sulphindigotate appeared to us so 
considerable that we have selected another pigment which is extremely easy to recognise 
in the smallest quantities, namely, fluorescin. We have also made experiments with 
three substances which are easily recognised by their colour reaction — sodium iodide, 
sodium ferrocyanide, and sodium salicylate. 

For these experiments we have used the rabbit, guinea-pig, cat, and dog. The two 


former are somewhat unsatisfactory, not only on account of their small size, but also from 
the fact that in both the development of the allantois is very incomplete. The same 
objection does not apply to the dog or cat, although in our hands the cat proved unsatis- 
factory, from the small amount of fluids in the later stages of gestation, when alone it 
was possible to be sure that the animals were pregnant. 

The method of procedure was to anaesthetise the animal, to fix a cannula connected 
with a burette containing the solution to be injected in the proximal end of the 
jugular vein, and to let the solution run in very slowly. In some experiments one 
cornu of the uterus, with its contained pregnant sacs, was removed to serve as a check, 
but in later experiments this was not done. 

When a sufficient quantity of the fluid had been run in — a good indication of this 
being given by the intense staining of the mucous membranes with the fluorescin — the 
injection was stopped and the animal allowed to live for from 10 to 20 minutes, being 
kept all the time under the anaesthetic. The uterus was then excised and the animal 
killed by pushing the anaesthetic. 

The uterus was at once removed to a clean table, and each sac was carefully opened 
and the fluids collected in carefully cleaned vessels, care being taken to prevent any 
admixture with the maternal blood. The foetuses were then removed, weighed, and if 
any urine was present in the bladder it was collected. Sometimes the foetal blood was 

The presence of fluorescin was recognised by its fluorescent appearance even in the 
most dilute solution ; and the amount was estimated by the intensity of the fluorescence. 

Salicylic acid was tested for with a dilute solution of ferric perchloride, and the 
depth of colour gave a fair index of the amount present. 

Sodium ferrocyanide was tested for with ferric chloride. 

Sodium iodide was tested for by the method given in the Appendix. 

The following experiment may be given in extenso as typical : — 

Experiment IX. — Irish terrier bitch, about 13 kilos., supposed to be seven weeks 


On December 14, anaesthetised at 10.30, and two sacs of the right cornu of the 

uterus at once excised. A cannula placed in the jugular vein and the following solution 

run in from a burette : — 

Sodium iodide ........ 10 grms. 

Sodium salicylate ........ 5 gnus. 

Fluorescin . . . . . . . . 0"5 grms. 

Water f . 100 

The injection began at 11.12, and at 11.42, 150 c.c. had been injected (i.e. 15 grms. of 
sodium iodide, 7 "5 grms. sodium salicylate, 075 grm. fluorescin). The dog was bled to 
death at 11.50. Eyes strongly fluorescent; skin and all the organs deeply stained. 
Blood-serum, bile, and urine intensely fluorescent. 








The two sacs removed at the beginning of the 

experiment contained : — 









The six sacs removed after death contained : — 


















91 36 



Foetus left in sac and fixed in formalin. 


Per cent. 










None of the fluids were fluorescent except those of 7, which were distinctly so. 

The allantoic and amniotic fluids of 4 and 7 gave a slight but distinct reaction of 
salicylic acid with ferric perchloride. 

The combined urine collected from the bladders was strongly fluorescent, but the 
urine of 6 was free of fluorescence. The combined urine also gave a strong reaction 
of salicylic acid with ferric perchloride. 

The foetuses, the allantoic and amniotic fluids, and about 1 of the combined 
urine were sent to Mr Kerr, who reports as follows, as regards the presence of Nal : — 

Presence of Nal. 

Foetuses (1 and 2 check), 142 grms. 

(3 to 7), 223 grms. 
Foetal urine, 2 c.c. 
Allantoic fluid, 140 c.c. 
Amniotic fluid, 60 c.c. 

Experiment I. — Guinea-pig, 570 grms., nearly full time. Injected 0'29 grm. sodium 
ferrocyanide in 4|- minutes. Killed after 15 minutes. Urine and kidney of mother 
gave marked reaction. Foetal fluid and foetal tissues gave no reaction. Maternal 
placenta gave a green colour. 

Experiment II. — Cat, about 2000 grms., nearly full time. 0*8 grm. sodium 
ferrocyanide in 13 minutes. Animal killed after 20 minutes. Kidney and urine of 
mother gave very marked reaction. Fluids, tissues, and urine of foetuses gave no 

Experiment III. — Rabbit, about 24 days. Sodium sulphindigotate (not fully 
saturated), 38 c.c. injected in 11 minutes. The maternal urine was blue- 
First sac removed 9 minutes after injection. 
Fluids not coloured. 
Urine ,, 




Second sac, 19 minutes after injection. 

Fluid faint green. 

Urine not coloured. 

Kidney ,, 
Third sac, 34 minutes after injection. 

Fluid not coloured. 

Urine ,, 

Kidney „ 

When mother killed, urine in bladder was green, but the kidneys were not coloured. 

Experiment IV. — Rabbit, about 20 days. Fluorescin, - 5 per cent., 10 c.c. in 6 
minutes. Five minutes after, abdomen opened ; abundant fluorescent ascitic exudation 
found. Ten minutes after, one cornu of uterus with three sacs excised. Membranes 
appeared fluorescent, but fluid was not so. The foetal blood was not fluorescent. 
Fifteen minutes after, the other cornu was removed, and gave the same characters. 
The urine and tissue of the mother were strongly fluorescent. 

Experiment V. — Guinea-pig, about 650 grms. 20 c.c. of - 5 per cent, fluorescin in 8 
minutes. Animal killed after 1 5 minutes. All the maternal tissues markedly fluorescent. 
No colour in fluids or in tissues of foetuses. 

Experiment VI. — Rabbit, almost full time. 20 c.c. 10 per cent, sodium iodide in 
12 minutes. Animal killed 13 minutes after. Two sacs, one with exomphalotic and 
anencephalic foetus, other with normal foetus but no fluids. 

Experiment VII. — Rabbit, 12-14 days. One sac excised as control. 20 c.c. sodium 
iodide (10 per cent.) in 13 minutes. Killed 12 minutes after. Amniotic fluid of six sacs 
mixed — also allantoic. Report : — 

Maternal urine 
Amniotic fluid 
Allantoic fluid 
Control foetus . 
Foetuses (6) after injection 

Presence of Iodide. 
strongly marked, 

distinct trace, 
strongly marked. 

Experiment VIII. — Rabbit, 17 days. Control sac excised. 20 c.c sodium iodide 
(10 per cent.) in 9 minutes. Animal died 6 minutes after injection stopped. The 
fluids of the remaining six sacs were mixed. Report : — 

Maternal urine 
Mixed fluids after injection 
Foetus .... 
Control fluids . 
Control foetus . 

Presence of Iodide, 
very strongly marked, 
distinct trace, 
strongly marked, 

Experiment IX. — Already described in detail. 

Experiment X. — Guinea-pig, nearly full time. One foetus of 92 grms. 20 c.c. of 
1 per cent, fluorescin in 10 per cent, sodium iodide in 5 minutes. Animal killed 10 



minutes after. Maternal blood fluorescent, but urine not so. Fluids not distinctly 
fluorescent. Foetal blood distinctly fluorescent. Report : — 


Presence of Iodide.* 

Maternal blood 

. - 0-026 grm. 

Foetus ..... 


Fluids . . . . . 


Foetal blood .... 

0-002 grm. 

Maternal urine 

0-0003 grm. 

Experiment XL — Guinea-pig, nearly full time. Two foetuses about 80 grms. 
1 7 "5 c.c. of same solution in 5 minutes. Animal killed after 10 minutes. Maternal 
blood fluorescent. Foetal blood fluorescent. Fluids slightly fluorescent. Report : — 

Sample. Presence of Iodide.* 

Maternal blood ........ 0-012 grm. 

Foetuses 0-002 grm. 

Foetal blood . . . . . . . . . present. 

Fluids 0-00006 grm. 

Experiment XII. — Cat. 45 c.c. solution of sodium iodide 10 per cent., sodium 
salicylate 5 per cent., sodium sulpliindigotate to saturation, were injected in 23 minutes. 
The animal died. Mucous membranes, all connective tissues, and muscles intensely 
blue. Medulla of kidney blue, cortex green. Liver and spleen not blue ; but when 
cut and exposed to air the surface of liver became greenish blue. Intestines deep blue, 
but when left covered by other coils lost their blue colour. Urine intensely blue, with 
red shimmer. Wall of uterus deep blue. Blood-plasma deep blue. 

Four sacs. Foetuses almost full time, weighing about 75 grms. each. 

Foetal organs and blood gave no trace of blue. 

Foetal urine, about 1 c.c, gave no trace of blue, but one drop gave a faint brown with 
FeCl 3 , indicating the presence of salicylic acid. 

Allantoic fluid of brown colour. No blue tinge. Gave faint deepening of brown 
with FeCl 3 . 

Amniotic fluid. — No trace of blue. Slight brown with FeCl 3 . Report states 
that there is a considerable quantity of iodide in the maternal urine (17 c.c.) and 
in the maternal blood (56 c.c), but only a trace in the allantoic fluid (7 c.c.) and in 
the amniotic fluids (13 c.c), and a trace in the four foetuses (300 grms.) = 0*0012 grm. 

Twelve experiments have thus been performed — five on rabbits, four on guinea-pigs, 
two on cats, and one on a dog. 

Sodium ferrocyanide has been injected twice, sodium sulphindigotate twice, sodium 
iodide seven times, fluorescin five times, and sodium salicylate twice. 

The following table gives the general result of these experiments, arranged under 

* In these two experiments the percentage of iodide present cannot be given, as the weights of the specimens 
were not noted. 





the different substances injected. The intensity of the reaction is indicated by the signs 
0, — , + , + + , and + + + indicating the most intense reaction. 

Substance Used and 
Species of Animal. 









Sodium Iodide. 


,, ... 
Rabbit .... 

Dog ... . 

Am. All. - 

+ + 

+ + 
+ + 
+ + 



+ + 
+ + + 




Am. + All. 

)) — >) — 

Sodium Ferrocyanide. 


Cat ... . 





Rabbit .... 

Dog ... 

Am. - ? All. - ? 




+ + 



+ + 

Sodium Salicylate. 

Rabbit .... 
Rabbit .... 
Dog ... . 

Am. -"All. - 





+ + 


Sodium Sulphindigotate. 

Rabbit .... 
Cat ... . 

Am. All. 




These experiments show (l) that, with the one very doubtful exception of sodium 
sulphindigotate in one sac in the rabbit in Experiment III., in no case did the substance 
injected into the maternal circulation appear in the fluids without also being present in 
the foetus ; and (2) that, when the substance appeared in the fluids, its amount in the 
fcetal tissues or foetal urine was always much greater. The unsatisfactory nature of 
sodium sulphindigotate for such investigations is well shown by Experiment XII. 

They therefore show that the transference of material from the mother to the fluids 
takes place through thefxtus. 

D. J. Copper (Der Ubergang bestimmtes Stoffe von der Mutter in das Fruchtwasser 
und in den Fotus. Utrecht, 1905) gives the results of three quantitative determinations 


of iodides in the fluids and in the foetal urine after their injection into two pregnant 
cows and one pregnant goat. In the two cows he finds the iodide in greater percentage 
amount in the amniotic fluid than in the urine ; while in the goat he finds it in 
about the same amount in the amniotic, but to a less amount in the allantoic than 
in the blood of the foetus. 

The very consistent results which we obtained with a series of chemical substances, 
in our opinion, counterbalances the value of these two observations, especially as the 
method employed by us in the determination of the amount of iodine present is more 
satisfactory than that used by Copper. 

But Copper does not attempt to argue in favour of the direct passage of the iodide 
from the maternal blood, but he believes that it passes through the maternal placenta to 
the chorionic villi of the foetal placenta, and that from the efferent vessels of this it is 
carried into the wall of the allantoic vesicle, which surrounds the amnion, and that from 
there it passes into the fluids. 

Gusserow (Arch. f. Gyn., Bd. iii. p. 241, 1872) had already indicated the possible 
importance of the vessels running over the placenta. 

Copper has repeated upon ruminants the experiments previously performed by 
ToERNGREN(C.i£. Soc. biol., 1888, p. 543) on rabbits, of injecting iodides into the amniotic 
sac and recovering them in the maternal urine. Such observations seem to prove no 
more than that iodides are taken up by the blood-vessels in the walls of the allantois 
and so carried to the placenta to be passed into the maternal circulation. They do not 
in any way indicate that these fluids are formed from the maternal blood-vessels. # 

2. The Mode of Production of the Fluids by the Foetus. 

It having been shown that the fluids and their constituents are foetal in origin, 
the question of how they are produced has next to be considered. 

(1) They may be produced as a transudation from the blood-vessels coursing in 
the walls of the vesicles, which, as has been indicated, are purely foetal in nature. 

(2) They may be derived from the kidneys of the foetus. 

(3) They may be derived from both these sources. 

(4) One fluid may be formed from the other to a greater or less extent, or 
each may derive constituents from the other by diffusion or osmosis through the 
amniotic and allantoic membranes which separate them. 

The question of the parts played by (1) transudation from the blood-vessels of the 
allantois and (2) by the action of the foetal kidneys may first be considered. 

It must be at once admitted that the small amount of fluid which is included in 
the amniotic sac at the time of its development must have an origin independent of 
blood-vessels and kidneys, and must be derived from a breaking-down or secretory 

* Copper gives a very excellent summary of the previous work on the subject. 


action of the cells of the ovum. With this earliest production of the fluid we are 
not concerned. 

In the sheep, the earliest pregnancies examined by us contained embryos of about 
1"5 grm. in weight, and in these the kidneys and the allantoic blood-vessels are both 
well developed and the urethra differentiated, while there is something less than 10 c.c. 
of amniotic and about 50 c.c. of allantoic fluid. 

Before the end of pregnancy these fluids had increased, the amniotic to an average 
of about 560 c.c, and the allantoic to 90 or 100 c.c. 

In the rabbit, the allantoic circulation and the development of the kidney are 
completed at about the ninth day. 

From an early stage of pregnancy both the factors under consideration are therefore 
capable of playing a part in the formation of these fluids, and the evidence as to the 
part played by each may now be examined. 

(1) The Evidence afforded by the Study of the Molecular Concentration of the 
Amniotic and Allantoic Fluids and of the Maternal and Fatal Blood. — A good many 
observations have been carried out along these lines by means of the cryoscope, in the hope 
of determining whether the fluids are secretions or transudates. The most satisfactory 
are those by Jacque and those by GtRUNBAUM. 

Jacque (Bull, de VAcad. de Belgique, 1902, p. 218) has examined the allantoic and 
amniotic fluids of the sheep cryoscopically in order to determine their molecular 
concentration. This he found always inferior to that of the blood, either foetal or 
maternal — allantoic fluid, A 0'522 ; amniotic fluid, A 0*538 ; maternal blood, A 0*578 ; 
foetal blood, A 0'623. From this he concludes that the origin of the fluids is not by 
transudation but by renal secretion. Further, by observations carried out at different 
periods of gestation, he finds that the molecular concentration of the two fluids varies. 
Thus, during the first stages of development, when the allantoic sac communicates freely 
with the bladder of the foetus, the concentration of the allantoic fluid is lower than 
that of the foetal urine, and lower than that of the amniotic fluid. At a later period, 
when the urachus is still patent and the urethra communicates with the amniotic cavity, 
the liquor amnii and the liquor allantoidis are of the same concentration ; while in the 
later stages of development, when the urachus is closed but the urethra still com- 
municates with the amniotic cavity, the amniotic fluid has a higher molecular con- 
centration than the allantoic. From this he concludes that in the early stages 
the allantoic fluid is derived from the foetal kidneys through the urachus, and that 
the amniotic fluid is derived from it by the abstraction of water. In the later 
stages the amniotic fluid is derived directly from the foetal kidneys by the urethra, 
and the allantoic is derived from the amniotic by a passage of water back through 
the membranes. 

Less satisfactory cryoscopic investigations have been carried out by Keim (Bullet, 
de la Societe d' Obstetrique, 1901), and by Billard, Dieulafe, and Gilles (Comp. 
Rend, de la Soc. de Biolog., 1905), and they conclude that in the early stages of 


development there is a current from the maternal blood-serum to the amnion, and in the 
later stages a current in the reverse direction from amnion to maternal blood-serum. 

Grunbaum (Verhand. de physic, med. Gesellsch. zu Wilrzburg, Bd. xxvii., No. 3, 
p. 161, 1905; abstract Malys Jahresbericht for 1905) found that the amniotic fluid 
had a lower molecular concentration than the maternal blood. The freezing-point 
reduction of the blood is - 53, of the foetal fluids 0'485, and of the foetal urine 0'2. 

In the cow, the molecular concentration of the amniotic fluid is nearly that of the 
blood till the end of pregnancy, when it becomes less. In the allantoic, the molecular 
concentration is at first very little different from that of the blood, later distinctly less, 
and towards the end of pregnancy again becoming nearly the same as the blood. So 
far as they were investigated, the dog, cat, and goat gave the same results as the cow. 
In the pig, the amniotic fluid shows a lower concentration than the blood in the second 
half of pregnancy, and the allantoic a lesser concentration. From these observations he 
concludes that in man the amniotic fluid is a mixture of transudation and fcetal urine, 
and that in the cow the amniotic fluid is entirely a transudation, the allantoic fluid in 
the beginning a transudation and later chiefly fcetal urine. 

The author appears to conclude that, when molecular concentration of one of the 
fcetal fluids approaches that of the blood, the fluid is to be regarded as a transudation, 
while if the molecular concentration is lower an admixture with urine is indicated. 

Considering the enormous variations in the relationship between the molecular 
concentration of the urine and blood in the adult, such a conclusion appears hardly 
warranted. While normally the A of the blood is 0"56° C, and that of the urine 2'3, the 
A of the urine may in the caffeine diuresis or after drinking large quantities of beer 
fall as low as '160. 

Again, the study of the absorption of fluids from serous spaces has shown that the 
molecular concentration of an effusion may be raised by the absorption from it of water. 
Thus, alterations in the molecular concentration of the fcetal fluids may be capable of 
explanation in terms of the relative activity of formation on the one hand and absorp- 
tion on the other. 

The conclusions arrived at by the study of the molecular concentration of the fluids 
have thus a somewhat unsatisfactory basis. 

It appeared to us that more light might be thrown upon the question of the relative 
part played by transudation on the one hand, and secretion by the kidneys on the 
other, by a more careful analysis of the results of our experiments on the injection of 
substances into the maternal blood, and a study of the chemistry of the fluids. 

(2) Evidence from the injection of Substances into the Maternal Blood. — The 
most convincing experiment is No. IX., on a bitch. Here fluorescin, sodium salicylate, 
and sodium iodide were injected. The combined urines of the foetuses were very 
strongly fluorescent, but the fluids of only one pregnancy sac showed a slight 
fluorescence. The fcetal urines gave strongly the reaction for salicylic acid, but the 


combined fluids gave only a faint although a distinct reaction. For the sodium iodide 

the figures show : — 

Foetal urine .......... 04 per mille. 

Allantoic fluid -0057 „ 

Amniotic fluid ......... *0066 „ 

There was nearly ten times as much in the foetal urine as in the fluids. Such 
observations seem to indicate that secretion from the kidneys alone would explain the 
appearance of these substances in the fluids. 

(3) Chemical Examination of Fluids. — The marked differences in the chemical 
constitution of urine and of exudations from blood-vessels justified us in hoping that 
a study of the chemistry of the fluids would throw important light upon their nature. 
Although a good many isolated observations on the chemistry of these fluids have been 
recorded, few systematic studies of the subject have been made. The best of these is by 
Doderlein {Arch. f. Gyn., xxxvii., Heft i. p. 141, 1890), who carried out an extended 
series of observations on the amount, physical characters, and chemical constitution of the 
fluids in the cow throughout pregnancy. Curiously enough, he failed to observe some 
important points, and, in our opinion, failed also to appreciate the significance of his 
own evidnce. 

Kestiakowski, in La Physiologiste Russe, 1898-99, p. 155, gives a series of analyses 
of the amniotic fluid of the cow at different stages of pregnancy, and he concludes that 
in the early part of pregnancy it is derived from the vessels of the placenta, and that 
in the later months it is an excretion from the organs of digestion. His analyses did 
not include the investigation of the non-protein nitrogen, and he thus failed to under- 
stand the significance of the presence of the urinary constituents in the fluids. 

For our investigations we have chiefly used the sheep, because (l) it is easy to get 
a considerable number of pregnant uteruses from the slaughter-house ; (2) in the sheep 
the allantoic and amniotic sacs are both well developed. 

More recently these observations have been extended by the examination of the 
fluids of the cow, pig, and dog. 

The results may be considered under the following heads : — 

(a) Amount, character, and specific gravity of each fluid at different stages of 

(b) The occurrence or non -occurrence of proteins in the fluids of various animals. 

(c) The amount, character, and distribution of the non-protein nitrogen. 

(d) The pigments. 

(c) The occurrence of sugar ; its nature and amount, and its relationship to the 
non-protein nitrogen. 

(/) The ash of the fluids. 

(a) Amount and Specific Gravity of the Fluids. — The following tables give the 
results of our observations during 1905-6 on the sheep. Table A gives the individual 
observation. Table B gives the average result at different stages of pregnancy. 



Quantity of Amniotic and Allantoic Fluids, and their Specific Gravity. 

Amniotic Fluid. 

Allantoic Fluid. 

Weight of 
Foetus, grms. 


Sp. Gr. 


Sp. Gr. 















• ■ • 


























• • • 




• • • 





































































14 + 



1 01 1 












































































TABLE A. — continued. 

Amniotic Fluid. 


c Fluid. 

Weight of 
Foetus, grms. 


Sp. Gr. 


Sp. Gr. 





















• ** 































Average Amounts op Amniotic and Allantoic Fluids, and their Specific Gravity, at Different 

Stages of Pregnancy. 



Weight of 
Foetus, grms. 

No. of Obser- 


Sp. Gr. 

No. of Obser- 


Sp. Gr. 

Below 15 







15 to 30 







30 to 100 







100 to 300 







300 to 1000 







Over 1000 







Doderlein (loc. cit., p. 149) found in the cow a marked and rapid increase in 
the amniotic fluid in the first half of pregnancy, and then a very marked fall 
towards the end ; and an almost continuous increase in the allantoic fluid — an increase 
which became more marked in the second half of pregnancy, when the amniotic fluid 
was decreasing. 

We have been unable to procure the original papers of Majewski (De substantia?* um 
quse liquor ibus amnii et allantoidis insinit: Dissertatio, Dorpat, 1858) on herbivora, 
or those of Tschernow (De liquorum embryonalicum in animalibus carnivoris : 
Dissertatio, Dorpat, 1858), referred to by Doderlein, p. 145. 



Our own few observations on the cow support Doderlein's results, as is shown in the 
appended table (C) : — 


Weight of Foetus, 




Present Obser- 


Present Obser- 


About 100 
600 to 700 . 
3000 to 4000 . 
About 9000 . 










In the pig we have made seven observations, which give the following average 
amounts of fluid in each gestation sac with different lengths of embryo : — 

Length of Embryo 

Amniotic Fluid 

Allantoic Fluid 

in cms. 

m c.cms. 

in c.cms. 






















This shows the same increase in the amniotic fluids as occurs in the sheep, with the 
same comparatively small increase in the amount of allantoic fluid. 

Amount. — In the animals examined, the amniotic fluid is at first small in amount, 
but rapidly increases, and in the sheep and cow reaches its maximum about the middle 
of pregnancy, after which it undergoes no very marked change in the sheep, but very 
markedly decreases in the cow. The allantoic fluid in the earlier part of pregnancy is 
larger in amount than the amniotic, but in the later stages it may show hardly any 
increase, as in the sheep, or it may increase enormously, as in the cow. 

Possibly these different relations depend upon the extent to which the amniotic 
fluid is swallowed by the foetus and re-excreted through the patent urachus into the 
allantoic vesicle. It is further possible that, in the middle period of pregnancy, before 
the muscular structures round the urethra are formed, the urine secreted passes equally 
into both sacs, but that in the later months in some animals, e.g. the cow, it may gener- 
ally pass into the allantoic sac, and only occasionally by micturition into the amniotic. 

Specific Gravity. — The allantoic fluid shows a marked increase in specific gravity. 
The amniotic fluid shows hardly any increase in its specific gravity. 





(b) Proteins of the Fluids. — Transudations from blood-vessels contain the proteins 
of the blood-plasma. In determining the source of the fluids, it was therefore necessary 
to examine them for the presence or absence of such proteins. 

The qualitative examination was made by acidifying and boiling for the detection 
of albumins and globulins, and by the addition of Esbach's solution for the detection of 
lower proteins. 

The following table gives the results : — 




Sheep (61) 


Cow (9) 



Pig (7) 



Cat (2) 



Dog (1) 


Kestiakowski records the presence of albumin in the amniotic fluid of the cow 
during the early stages of pregnancy only, and the presence of mucin in the later stages. 

The variations in the amount of proteins in the allantoic fluid throughout pregnancy 
were studied in the sheep by precipitating either by boiling with addition of acetic acid, 
or by the addition of an equal quantity of a 10 per cent solution of trichloracetic acid, 
and weighing. 

The following table gives the results : — 


Weight of Foetus 

Per cent, of Protein 


in grms. 

in Allantoic Fluid. 
























150 to 650 




400 to 800 












They show that the amount of protein varies little throughout the greater part of 

According to Doderlein's observations on the cow, the amount of proteins in both 
fluids rises throughout pregnancy in the amniotic from '042 to "45, and in the allantoic 
from 0-3 to V37. 

In the dog, six weeks pregnant with embryos of 35 grms., the allantoic fluid was free 
of proteins, but the amniotic contained distinct amounts of native proteins amounting to 
*06 per cent. 

In another dog, said to be seven weeks pregnant, and with foetuses averaging 100 
grms. in weight, a mere trace of protein was found in the allantoic fluid, while a very 
distinct amount was present in the amniotic. 

In the pig's uterus at various stages of pregnancy both amniotic and allantoic 
fluids contained native proteins, the former 0*495 per cent., the latter 0*563 percent. 

The proportion of albumin and globulin was determined in four cases in the sheep : — 

Proportion of Albumin and Globulin in Allantoic Fluid. 

Numbers of 
Fluids mixed. 

"Weight of Foetuses, 

Total Protein. 

Per cent, of 





40 to 50 

300 to 500 

500 to 600 







The fact that coagulable proteins do not exist in the amniotic fluid of the sheep 
seems to oppose the view that it is formed as a transudation from blood-vessels, because, 
so far as we at present know, all such transudations contain the proteins of the blood- 

On the other hand, their presence in the allantoic fluid of the sheep, and in both 
fluids in the other animals observed, does not oppose the view that these fluids are 
formed in the kidneys. 

The opportunity of examining foetal urine is not often afforded, because the bladder 
is generally empty. Doderlein (op. cit., p. 155) gives two analyses, one of the urine of a 
foetus of 15,000 grms. and another of a newly-born calf. In the first there was 0'33 per 
cent., in the second 0*18 per cent., of proteins. In the urine of two foetal sheep, one of 
1505 grms. and one of 1630 grms. — nearly full time — we found coagulable proteins 
present. This is confirmatory of the old observations of Dzardi in 1806, that alcohol 
causes a white gelatinous precipitate in the urine of the new-born calf, and of those of 
Virchow, Martin Ruge and Biedermann, and of Dohrn in the new-born child (quoted 
by Doderlein, loc. cit., p. 172). Jacque (loc. cit.) also records the presence of albumin 
in the urine of foetal sheep. 



(c) Non-Protein Nitrogen and its Distribution. — While the absence of native proteins 
affords evidence of some value against the origin of the fluids by transudation, and while 
their presence does not militate against the idea that the fluids are of renal origin, a study 
of the presence of non-protein nitrogen, and the nature of the compounds in which it 
occurs, should throw light on the relation of the fluids to foetal urine. 

The total nitrogen was determined in 5 or 10 c.c. by Kjeldahl's method. In the 
allantoic fluids the total proteins were determined by precipitation with trichloracetic 
acid, and the nitrogen either estimated directly or calculated from the weight of the pre- 
cipitate. This was subtracted from the total nitrogen, and the non-protein nitrogen was 
thus obtained. 

The urea nitrogen was determined generally by the method of Moener and Sjoquist, 
sometimes also by the method of Bohland. By the former method a considerable 
quantity of any allantoin present is estimated with the urea, and by the latter method 
all is probably so estimated. The results of the earlier analyses were rejected, as the 
amount of fluid taken (5 c.c.) proved to be too small to yield an adequate amount of 
ammonia for accurate titration. 

Generally the urea results by Bohland's method were considerably higher than those 
by Morner's, especially in the case of the allantoic fluids. 


Distribution* of Nitrogen in Amniotic and Allantoic Fluids. 




Per cent, of 



M. and S. 





M. and S. 

M. and S. to 

of Foetus. 


































• • • 


















































































• • • 
















TABLE G— continued. 



Per cent, of 



M. and S. 





M. and S. 

M. and S. to 

of Foetus. 






























! 720 




















j 1670 














Percentage of Non-Protein Nitrogen in Amniotic and Allantoic Fluids at Different Stages 

of Pregnancy. 

Weight of Foetus, grms. 



No. of 

Per cent, of 

No. of 

Per cent, of 

Below 15 

30 to 100 . 

100 to 300 . . . 

300 to 1000 

Over 1000 . 









Percentage of M. and S. Nitrogen in Amniotic and Allantoic Fluids at Different Stages of 
Pregnancy, and its Relation to the Total Non-Protein Nitrogen. 

Weight of Foetus, 



Per cent, of 

Per cent, of 


No. of 

M. and S. 

M. and S. N. 

No. of 

M. and S. 

M. and S. N. 



to Total Non- 
Protein N. 



to Total Non- 
Protein N. 

Below 15 







30 to 100 







100 to 300 . 







300 to 1000 . 







Over 1000 








The results of our determination of the non-protein nitrogen show that excretory 
products are constantly present in both fluids throughout the whole of pregnancy, 
and that in the sheep, as Doderlein found in the cow, an increase occurs as pregnancy 
advances. From the beginning and throughout the whole time the amount is higher 
in the allantoic than in the amniotic fluid. 

The study of the distribution of nitrogen in the various excretory substances is still 
under investigation, but the observations so far made show — 

1st. That in the amniotic fluid somewhere between 70 and 80 per cent, of the 
nitrogen is in the form of urea (or allantoin?), and that the proportion is not altered 
throughout the course of pregnancy. 

2nd. That in the allantoic fluid only about 50 to 70 per cent, of the nitrogen is in 
urea and allantoin. 

That a large amount of this nitrogen of the amniotic fluid is urea has been demon- 
strated by the separation of urea as the oxalate. 

The constant occurrence of nitrogen in simple combinations, and the fact that a 
considerable proportion of this is urea, is a strong argument in favour of the view that 
both allantoic and amniotic fluids are derived from the foetal kidney. 

(d) Pigment. — At the beginning of pregnancy both allantoic and amniotic fluids 
are colourless, the amniotic fluid being water-clear and the allantoic slightly turbid ; 
but in the later stages the amniotic fluid of the sheep becomes of an amber-yellow 
colour. In the cow the allantoic fluid has the same colour, but the amniotic fluid often 
becomes colourless. 

After shaking with ether a brown fluid is left, which gives the spectrum of urobilin. 
This further indicates the renal origin of the fluid. 

(e) Sugar. — The earliest record of the presence of sugar in the foetal fluids which we 
have found is by Majewski (loc. cit., 1858). Subsequent observers, e.g. Doderlein, 
have failed to observe it. 

We have found it to be a constant constituent of both allantoic and amniotic fluids 
of the sheep, cow, pig, cat, and dog, and of the mixed fluids of the rabbit, ferret, and 

The variations in its amount in the two fluids throughout pregnancy have been 
studied in the sheep. The analyses were made in the usual manner by Fehling's 
method. The results are given in the following tables (K and L) : — 




Distribution of Sugar in Amniotic and Allantoic Fluids. 

Weight of Foetus, Amnio 

tic Fluid. 

Allantoic Fluid. 

grms. Per cen 

-j. of sugar. 

Per cent, of sugar. 





























• • • 














• • •' 


























Percentage of Sugar in Amniotic and Allantoic Fluids at Different Stages of Pregnancy. 

Weight of Foetus, i No. of Obser- 
grms. vations. 

Amniotic Fluid. 
Per cent. 

No. of Obser- 

Allantoic Fluid. 
Per cent. 

Below 15 

30 to 100 

100 to 300 

300 to 1000 

Over 1000 









In the allantoic fluid the amount varies little, being "409 per cent, in the earlier 
pregnancies, '438 per cent, in the later stages. 

In the amniotic fluid the percentage is at first small — only "083 — and it rises in the 
later period to '287 — an increase of more than threefold ; but it is always smaller than 
the percentage in the allantoic, except possibly at the very end of gestation. 

In a cow about the middle of pregnancy the percentage was : — 


L00 per cent. 
0-277 „ 



Nature of the Sugar. — The sugar in the amniotic and allantoic fluids of the sheep 
and cow reduces Fehling's solution, ferments with yeast, and yields crystals with the 
character of glucosazone. But it is laevo-rotatory and gives Seliwanoff's lsevulose 
reaction. It is therefore the ketose, lsevulose. In the allantoic fluid of the dog, the 
sugar was in such small amounts that its nature could not be investigated. 

G-urber and Grunbaum (Munch, med. Wochen., 1904, 377) have already recorded 
the presence of leevulose in these fluids in the sheep, cow, and goat. 

Source of the Sugar. — The following facts indicate that the sugar is derived from 
the foetus : — 

1st. We have found the same sugar in the foetal urine of the sheep to the extent of 
0"5 to 0*7 per cent. 

2nd. It is present in the foetal blood. By extracting large quantities of foetal blood 
with alcohol, evaporating with charcoal, and taking up in a small quantity of water, we 
were able to extract a sugar which was lsevo-rotatory and which gave Seliwanoff's reaction 
distinctly. The fact that it occurs in larger amounts in the urine than in the blood 
indicates that its presence is not simply due to the swallowing of amniotic fluid by the 
foetus. In one case the amount of sugar in the blood was as much as 0'42 per cent. 

We have been unable to demonstrate this sugar in the liver. The amount of 
carbohydrate present in that organ in the foetal lamb seems to be very small. Chipman 


Nitrogen-Sugar Relations. 



Weight of 

Foetus, grms. 

Non-Protein S 






Nitrogen. ( 

















































• • ■ 





• • • 

• *• 





• ■ • 




































2 ; 8 





















{Reports from the Laboratory of the Royal College of Physicians of Edinburgh, 
vol. viii., 1903) found glycogen absent from the foetal rabbit on the 22nd day, but 
present in the placenta from the 8th day. 

The demonstration of the origin of the Isevulose from the foetus and of its presence 
in such considerable quantities in the foetal urine is further proof that the foetal fluids 
are derived from the foetal kidneys. 

The presence of Isevulose as a normal constituent of the blood and urine of the foetus 
is of considerable interest in connection with the occurrence of cases of lsevulosuria in 
man without any apparent cause. 

The relationship of the sugar to the non-protein nitrogen throughout pregnancy is 
of interest. While in the amniotic fluid the sugar is increased threefold and the 
nitrogen undergoes a much smaller increase, in the allantoic fluid the nitrogen is 
increased more than threefold, while the sugar undergoes only a slight rise. 

The proportion between nitrogen and sugar thus varies as is shown in Tables M 
and N. 

Relationship of Sugar to Non-Protein Nitrogen at Different Stages of Pregnancy. 

Weight of Foetus, 





Below 15 . 
30 to 100 . 
100 to 300 
300 to 1670 . 





(/) Salts. — The salts of the foetal fluids in the cow throughout pregnancy have been 
investigated by Doderlein (loc. cit., p. 161 et seq.). He finds the chlorides are higher in 
amount in the amniotic than in the allantoic fluid, and that in both fluids the percentage 
varies little throughout pregnancy. The average, as NaCl, is in the amniotic fluid 
0*59 per cent., and in the allantoic 0"23. The sodium behaves like the chlorine. 
Potassium remains constant at about 0*05 per cent, in the amniotic fluid, but in the 
allantoic fluid it increases from - 03 to 0*1 3. He finds a very marked difference in 
the percentage of magnesium in the two fluids, it being almost absent from the 
amniotic, but present in the allantoic to the extent of 0'01 to 0*05 per cent. 

Our analyses give the following results. The methods used are given in the 






Inorganic Constituents op Amniotic and Allantoic Fluids. 




Size of Embryos . 



to 650 
















Amniotic Fluid. 

Volume sent, in . 
Grms. per 100 c.c. of fluid — 

CI . 

P0 4 . 

S0 4 . 

Ca . 

Mg . 

K . 

Na . 

























Allantoic Fluid. 
Volume sent 

CI . 
P0 4 . 

so 4 . 

Ca . 
Mg . 
K . 

Na . 






















Our results on the cow and sheep confirm the observations of Doderlein, as is shown 
in the accompanying table, in the top line of which the averages of all the figures given 
by Doderlein are shown, and in the second line the averages of our results are given — 






P0 4 . 

so 4 . 

I Dodbrlein's Results . 
Amniotic ' 

{ Present Results . 








| Poderlein's Results . 
Allantoic -| 

( Present Results . 








We further find — what Doderlein did not record — an excess of phosphorus and 
sulphur compounds in the allantoic fluid. Our single examination of the fluid in the 
pig does not show the differences in the magnesium and calcium contents of the fluids 
which are observed in ruminants. 

Doderlein lays great stress upon the distribution of these inorganic constituents, 
maintaining that the high content of the amniotic fluid in chlorine and sodium indicates 
its origin as a transudate from the plasma, and that the large amount of magnesium 
in the allantoic fluid indicates its origin from the kidneys. But Vierordt 

* These results should be lower, as in some cases he gives " traces." 


{Daten und Tabellen, p. 333) gives a table based upon analyses of Schmidt and 
Vogel showing a content of NaCl in the blood- plasma of "5 5 46 per cent., and in the 
urine of 1'10 per cent., while in flesh and many other constituents of the body the 
proportion of magnesium is greater than that of calcium. It seems unsatisfactory to 
attempt to draw conclusions as to the source of the fluids from their inorganic 

III. Summary of Evidence and Conclusions. 

To sum up the evidence at our disposal : — 

1. As against the fluids being of maternal origin, it has been shown : — 
1 st. That anatomically this is improbable. 

2nd. That pathological evidence gives no support to the idea. 

3rd. That, after the foetus has been killed by withdrawing the fluid, no re-formation 
of the amniotic fluid occurs, although the maternal uterus and placenta remain normal. 

4th. That various substances injected into the maternal vessels appear in the 
fluids only after first being in the foetus. 

The only evidence in favour of the maternal origin of the fluids is the experimental 
work of Zuntz and Wiener with sulphindigotate of soda on rabbits — the probable 
fallacies of which have been explained. 

2. As against either of the fluids being formed as a transudation from the foetal 
blood-vessels in the wall of the allantois may be adduced : — 

1st. The absence of the proteins of the blood-plasma from the amniotic fluid of the 
sheep and from the allantoic fluid of the dog. Their presence in the fluids of other 
animals cannot be considered as indicating their origin as transudations, since these 
proteins are present in foetal urine. 

2nd. The observations recorded on the molecular concentration of the fluids and 
on the changes which this undergoes throughout pregnancy afford no evidence of 
their formation as transudations. 

The only evidence of an exchange between these vessels and the fluids is the fact 
that sodium iodide is so readily absorbed from the fluids and passed through the 
placenta into the maternal circulation. But the absorption of such a substance into 
the blood-vessels does not prove that normally fluid passes from these vessels into 
the foetal fluids, any more than its absorption from the bladder proves that normally 
the urine is secreted from the bladder wall. 

3. That the fluids are of the nature of urine derived by secretion from the foetal 
kidney is indicated by the following facts : — 

1st. The anatomical connection of the two cavities with the foetal urinary system. 
2nd. Both fluids contain a considerable proportion of nitrogen in urea and allied 


3rd. The existence of a peculiar sugar — laevulose — in the fluids, and also in the 
foetal urine. 

4th. Proteins, which are generally, although not always, present in the fluids, have 
been found in the fcetal urine. 

5th. Gusserow observed that, when benzoic acid is injected into the mother, hip- 
puric acid appears in the fcetal fluids. 

6th. Such substances as sodium iodide, sodium salicylate, and fluorescin, which, 
when injected into the maternal circulation, appear in the fcetal fluids, are found in much 
greater abundance in the fcetal urine and tissues. 

All investigators who have recognised the foetal origin of the fluids admit that the 
allantoic fluid is at least mainly urinary in nature, but it has been maintained by most 
that the amniotic fluid is a transudation. 

Thus Ellenberger ( Vergleichende Physiologie der Haussaugthiere, Bd. ii. p. 523, 
1892) says : " The amniotic fluid is a transudation from the rich cutaneous vessels of the 
embryo ; later it is also possibly derived from the vessels of the allantois which surround 
the amnion, . . . into which, towards the end of pregnancy, urine and saliva may escape." 

Doderlein also regards the amniotic fluid as a transudation, basing his conclusion 
entirely upon the correspondence between the salts of the fluid and those of the blood- 
plasma, and ignoring his own results upon the non-protein nitrogen and the other 
evidence which is forthcoming. 

The last question which must be discussed is a very difficult one. Admitting the 
origin of both fluids as essentially a secretion from the fcetal kidneys, does any exchange 
go on between the fluids, separated as they are by only a thin double membrane '? 

The facts to be taken into consideration in dealing with this question seem to be 
the following. They are diagrammatically represented in fig. 2 : — 

1st. The relatively rapid increase of the amniotic fluid as compared with the allantoic. 

2nd. The steady increase of the sp. gr. of the allantoic fluid without any change 
in the sp. gr. of the amniotic, and the fact that in the early stages the sp. gr. of the 
two fluids is the same. 

3rd. The steady increase of non-protein nitrogen in the allantoic fluid, and the 
slighter increase of the nitrogen of the amniotic fluid. 

4th. The high proportion of urea nitrogen in the amniotic fluid and the steady 
decrease in the proportion of urea nitrogen in the allantoic fluid. 

5th. The steady increase of sugar in the amniotic fluid, so that at the end of 
pregnancy it becomes equal to, or greater than, the percentage in the allantoic, which 
from the beginning remains fairly steady. 

6th. The high chlorine content of the amniotic fluid as compared with the 
allantoic throughout pregnancy, and the low magnesium content of the same fluid. 

The higher chlorine and possibly the higher sodium content (Doderlein, Kerr), with 



the higher molecular concentration found both by Jacque and Grunbaum (fig. 1), of the 
amniotic fluid early in pregnancy might cause an osmosis of water from the allantoic to 
the amniotic sac (fig. 3). This might account for a concentration of the allantoic fluid 
and the rise in the sp. gr. and in the non-protein nitrogen ; and, if the amniotic fluid 






Fig. 1. — To show changes in the molecular concentration of the fluids throughout gestation. 

were also being added to from the kidneys, it might explain the slight change in the 
sp. gr. and in the non-protein nitrogen in the amniotic fluid. 

On the other hand, diffusion might explain the passage of sugar and of the nitrogen 
in the form of readily diffusible urea from the allantoic to the amniotic fluid, thus 

Allantois Amnion 

Foetal Bladder 



Fig. 3. 

— Amount 
-,- Sp. C. 

■-- Hon- protein N: 

— Sugar 

Fig. 2. 

raising the percentage of sugar in the amnion and maintaining the percentage of 
non-protein nitrogen (fig. 3). 

The absence of further increase in the amniotic fluid in the later stages of pregnancy 
may be due partly to the foetus swallowing the fluid, and, in the cow at least, passing it 
into the allantoic sac ; but it may in part be due to the rise in the molecular concentration 
of the allantoic fluid (Jacque and Grunbaum) arresting the osmotic stream. This would 


not interfere with the diffusion of sugar, and hence the percentage of that substance in 
the amnion would tend to rise till it equalled that in the allantoic fluid. But if such 
exchanges go on, it is difficult to explain the persistence of the high chlorine content of 
the amniotic fluid. Why should it not pass into the allantoic fluid ? Or is it possible 
that the passage of water and diffusible solids can take place only from allantois to 
amnion, but not in the reverse direction ? 

The whole question of these exchanges is further complicated by : — 

(a) The continuous addition of urine to the allantoic fluid. 

(b) The intermittent addition of urine to the amniotic fluid. 

(c) The possible absorption of allantoic or even of amniotic fluid by the blood-vessels 
going to the placenta. 

These various considerations do not affect our main thesis, that both foetal fluids are 
essentially derived from the foetus through the kidney. 

Our thanks are due to Mr A. M. Trotter, M.R.C.V.S., for his kindness in procuring 
for us a considerable amount of the material used in these investigations. 


The Method used for Detection and Estimation op Iodide. 

The process employed for the detection and approximate determination of the iodide 
was as follows : — 

The specimen was ignited to ash after addition of solid sodium carbonate. The ash 
was boiled with water, filtered, concentrated to definite volume. 

A measured volume of the solution was taken, such that it contained not more than 
0*0005 gramme of iodide calculated as sodium iodide, 1 c.c. of chloroform added, and 
1 c.c. of concentrated solution of potassium nitrite, and the mixture acidified with dilute 
sulphuric acid. 

The colour of the resulting solution of iodine in chloroform was then compared with 
the colour obtained by the same process from a standard solution of iodide. 

The smallest quantity of iodide which could be detected by this method was 
equivalent to '00003 6 gramme of sodium iodide. 

A blank experiment was conducted, and gave no indication of iodide. 

Determination of Chloride. 

A measured quantity of fluid was evaporated to dryness in a platinum basin, after 
addition of some pure solid sodium carbonate. 

The residue was carefully carbonised, digested with dilute nitric acid, and to the 
solution a measured excess of standard silver nitrate solution was added. 

The precipitated silver chloride was filtered off, and the excess of silver nitrate in 


the filtrate was determined by means of standard ammonium thiocyanate solution, using 
ferric sulphate solution as indicator. 

The difference between the amount of silver nitrate added and the amount found 
after precipitation of chloride was calculated into the equivalent amount of chlorine. 

Determination of Total Sulphur. 

The total sulphur was determined by igniting with sodium carbonate and potassium 
nitrate, dissolving the ash in dilute hydrochloric acid, and precipitating with barium 
chloride solution. 

The weight of barium sulphate obtained, multiplied by the factor 0'41146, gave the 
proportion of total sulphur expressed as S0 4 . 

Determination of Total Phosphorus. 

The total phosphorus was determined by igniting with sodium carbonate and 
potassium nitrate, dissolving the ash in dilute nitric acid, and precipitating the boiling- 
hot solution with a boiling-hot solution of ammonium molybdate in presence of 
ammonium nitrate and nitric acid. 

The precipitated ammonium phospho-molybdate was filtered, washed with a mixture 
of ammonium nitrate and nitric acid solutions, dissolved in ammonia, the solution 
heated to the boiling-point, and the ammonium phospho-molybdate re-precipitated by 
addition of boiling; -hot nitric acid solution. 

The precipitated ammonium phospho-molybdate was filtered through a Gooch's 
crucible, washed with ammonium nitrate and nitric acid solution, and heated at 160° C. 
until the weight was constant. 

The weight of ammonium phospho-molybdate found, multiplied by the factor 0*05022, 
gave the proportion of total phosphorus expressed as P0 4 . 

Determination of Calcium and Magnesium. 

The ash from a portion of the fluid was dissolved in dilute hydrochloric acid, 
filtered, warmed, made alkaline with ammonia, and then acidified with acetic acid. 

The calcium was precipitated by addition of ammonium oxalate, the mixture 
heated for some time, and then filtered. The precipitate was ignited in a platinum 
basin, then moistened with dilute sulphuric acid, dried, ignited, and weighed as 
calcium sulphate. 

The weight of calcium sulphate, multiplied by 0*29450, gave the proportion of 
calcium compounds expressed as Ca. 

In the filtrate from the calcium oxalate, the magnesium was precipitated by sodium 
phosphate and ammonia as magnesium ammonium phosphate. 

After standing twenty-four hours this was filtered off, washed with dilute ammonia 


solution, ignited in a porcelain crucible over a blowpipe, and weighed as magnesium 

This, multiplied by the factor 0*21875, gave the proportion of magnesium compounds 
expressed as Mg. 

Determination of Potassium and Sodium. 

A measured volume of the fluid was evaporated to dryness in a platinum basin and 
carefully ignited over a low flame. The ash was dissolved in dilute hydrochloric acid 
and filtered. The filter was ignited, dissolved in dilute hydrochloric acid, and the 
solution filtered. 

The mixed filtrates were made alkaline with barium hydroxide solution, warmed for 
some time and filtered. The filtrate was evaporated to small volume, and the excess of 
barium removed by repeated treatment with ammonium carbonate and filtration. 
Finally, the filtrate was acidified with hydrochloric acid, made alkaline with ammonia, 
and evaporated to dryness. The ammonium salts were then burnt off in a platinum 
basin at a low temperature, and the residue weighed. This residue was dissolved in 
water and filtered. The filter was washed, burnt, and weighed, and its weight subtracted 
from that of the undissolved residue. 

The resultant is the weight of the mixed alkali chlorides. 

The filtrate was treated with excess of hydrochloro-platinic acid, evaporated to 
dryness on water-bath, treated with 85 per cent, alcohol, and filtered. The precipitate 
was washed with 85 per cent, alcohol, and dissolved through the paper with boiling 
distilled water. The solution was collected in a small weighed beaker, in which it was 
carefully evaporated to dryness, dried at 130° C. and weighed. 

The weight of potassium chloro-platinate thus obtained, multiplied by the factor 
0"3056, gave the weight of potassium chloride. 

By deducting the weight of potassium chloride from the total weight of alkali 
chlorides obtained above, the proportion of sodium chloride present was determined. 

The weight of potassium chloride found, multiplied by the factor 0*52479, gave the 
proportion of potassium compounds expressed as K. 

The weight of sodium chloride found, multiplied by the factor 0*39401, gave the 
proportion of sodium compounds expressed as Na. 

( 103 ) 

IV. — Report on Fossil Fishes collected by the Geological Survey of Scotland 
from Shales exposed on the Shore near Gullane, East Lothian. By Ramsay 
H. Traquair, M.D., LL.D., F.K.S. (With Two Plates.) 

(Read May 6, 1907. MS. received May 25, 1907. Issued separately December 21, 1907.) 

In the year 1903 Mr Macconochie made an interesting collection of fossil fish- 
remains from beds of Lower Carboniferous age exposed on the seashore near Gullane, in 
East Lothian, which were submitted to me for determination by the Director of the 
Geological Survey. Accordingly, I prepared a preliminary account of these remains, 
which was published in the Summary of Progress for the year above mentioned. 

Since that time Mr Macconochie has collected additional specimens from the same 
locality, which has also been visited by several private collectors, to whom I am 
indebted for kindly permitting me to examine and record the specimens which they 
obtained. So, with the sanction of the Director, I propose to devote this paper to a 
detailed, as well as illustrated, account of the results obtained by the searching of the 
beds in this new Scottish locality for Carboniferous Fossil Fishes. 

The locality in question is situated on the shore about two miles north-east of 
Gullane, and three and a half miles straight west from the centre of the town of North 
Berwick. I am informed that locally the spot is known as " Cheese Bay," on account of 
a ship laden with cheese having been wrecked there. Here three beds lying pretty 
close together have yielded fossil fishes, and these, commencing with the lowest, may be 
noted as under. 

Bed a. A soft shale of about 2 feet in thickness, showing on cross fracture a fine 
banded coloration ; layers of a dark purplish grey about \ inch in thickness alternating 
with others of lesser depth and paler appearance ; but the proportions of these bands 
are not always the same, and frequently they are considerably thinner than the size 
given above. This is the bed which has yielded the greatest number of species of 
fishes, although to obtain any at all much patient labour is necessary. 

Bed b. Above this fish-bed lie several feet of shale, partly partaking of the nature 
of an oil-shale, and on the top of this is a band, the presence of fish-remains in which 
was first detected by Mr John Muir. The shale constituting this band is very fissile, 
weathering with ochreous stains, and containing many coprolites ; but the fishes which 
it contains are few in point of species and badly preserved. 

Bed c. Besting on this last-mentioned fish-bearing shale is a hard, ferruginous 
cement-stone band, about six inches in thickness, containing numerous specimens of 
small Crustacea such as Tealliocaris Loudonensis, Peach. A few fishes occur in this 
band, which is, however, much more productive in the Arthropoda referred to, and which 
are mostly in a very fine state of preservation. 



We may now proceed to the enumeration and, where necessary, the description of 
the fishes and fish-remains which have occurred in these beds. 

Class PISCES. 
Subclass Elasmobranchii. 


Family Acanthodid^e. 
Genus Acanthodes, Agassiz. 

Acauthodes, sp. indet. 

The collection submitted to me contains four fragmentary specimens referable to 
this genus, but of which it is not possible to define the species with accuracy. The best 
of these is a portion, l l\ inches in length, of the hinder extremity of a fish, showing the 
dorsal and anal spines, but cut off just at the origin of the caudal fin. The minute 
quadrate scales have mostly a small dimple in the centre of the outer surface. 

Though I hesitate to affix a specific name to these fragments, I rather suspect they 
will turn out to belong to the same species as the fragment from Wardie named by 
Agassiz A. sulcatus (Poiss. Foss., vol. ii., pt. 1, 1835, p. 125, pi. 1 c, figs. 1, 2). 

Subclass Teleostomi. 


Family Rhizodontid^e. 

Genus Rhizodus, Owen. 

Rhizodus Hibberti (Agassiz and Hibbert). 

1836. Meyaliclitltys Hibberti, Agassiz and Hibbert, Trans. Roy. Soc. Edinb., vol. xiii. p. 202, pi. viii. 

fig. 1 (non figs. 2 and 3), pi. ix. figs. 2, 3, 9, 10; woodcut on p. 183. 
1841. Rhizodus Hibberti, Owen, Odojitoyrapliy, vol. i. p. 75, vol. ii. p. 12, pi. xxxv. fig. 2, pis. xxxvi., 

1841. Ho/optyc/iius Hibberti, Agassiz, Poiss. Foss., vol. ii. pt. ii. p. 180 (name only). 
1855. Rhizodus Hibberti, M'Coy, Brit. Palseoz. Foss., p. 612. 

,, Rhizodus gracilis, M'Coy, ibid., p. 611, pi. iii. G, fig. 17. 
1877. Rhizodus Hibberti, R. H. Traquair, Proc. Roy. Soc. Edinb., vol. ix. p. 658. 
1891. ,, „ A. S. Woodward, Cat. Foss. Fishes Brit. Mus., pt. ii. p. 342. 

1899. Megalichthys Hibberti, O. P. Hay, American Natttralist, vol. xxxiii., 1899, p. 787. 

Two scales, one of which measures 3 inches in diameter, have been found by Mr 
Macconochie in bed a, and one large tooth 1\ inches in length, deficient at the point, 
by Mr W. T. Gordon in bed b. 


There can be no doubt that this is the fish to which Agassiz originally applied the 
term Megalichthys, in allusion to the great bulk which it must have attained, as testified 
to by the gigantic teeth and scales, which were the first of its relics to come into the 
hands of Dr Hibbert ; and there can also be no doubt that, according to the strict laws 
of nomenclature, Megalichthys is still its legitimate generic name. It is true that 
Agassiz confounded its remains with those of a rhombic-scaled Osteolepid, which he 
afterwards (1844) described and figured under the name of Megalichthys Hibberti, 
while erroneously relegating the original Megalichthys to the genus Holoptychius. 
But priority for the application of the name Megalichthys to the fish now under con- 
sideration was secured eight years previously by the fact that Hibbert, in pi. viii. of 
his well-known Burdiehouse memoir, quoted above, figured under that name one of its 
large teeth as fig. 1, while the rhombic scale of the Osteolepid appeared on the same 
plate as fig. 3. I am consequently of opinion that I was wrong when in 1884 * I said 
that, " though it may be a matter of regret that it "— the genus now known as Rhizodus 
— " did not retain the name Megalichthys, the laws of nomenclature do not admit of 
any alteration now " ; and that Mr 0. P. Hay is technically right in wishing to restore 
that naine to the "big fish" of the Scottish Lower Carboniferous period. But the 
inconvenience which would be caused at the present time by the replacement of 
Rhizodus by Megalichthys would be such that I prefer meanwhile to adhere to the 
current nomenclature. 

Family Ccelacanthid^:. 

Genus Ccelacanthus, Agassiz. 
Ccelacanthus, sp. indet. 

Since my preliminary note on the Gullane fishes was written, various fragments of 
a species of Ccelacanthus have occurred in the highest or " Crustacean " bed, but 
unfortunately none are sufficiently perfect to enable one to refer them to a known 
species, or to found a new one for their reception. The species must have attained a 
respectable size for a Ccelacanthus, as some of its scales are nearly § inch in antero- 
posterior diameter ; their exposed surface is covered with an ornament approximately 
similar to that in C. elegans, Newb. (lepturus, Ag.), but to all appearance finer and 
more delicate. The sculpture of the head bones is not shown in any of the specimens, 
so that no satisfactory comparison with other species can as yet be obtained. 

Bed c. Collected by Mr Macconochie, and also found by Messrs R. Dunlop and 
Angus Peach. 

*Geol. Mag. [3], vol. i., p. 115. 



Family Pal^oniscid^;. 

Genus Gonatodus, Traquair. 

Gonatodus punctatus, Agassiz, sp. PL IT. figs. 1,*2. 

1835. Amblypterus punctatus, Agass., Poiss. Foss., vol. ii. part i. p. 109, pi. iv.c fig. 4 (non figs. 3, 5-8). 
1S72. Amblypterus anconooschmodus, R. Walker, Trans. Edinb. Geol. Soc, vol. ii. pt. i. p. 119, 

with plate. 
1877. Gonatodus punctatus, R. H. Traquair, Qu. Journ. Geol. Soc, vol. xxxiii. p. 555, and Proc. Roy. 

Soc. Edinb., vol. ix. p. 265 ; and Ganoid Fishes Brit. Carb. Form. (Pal. Soc), pi. ii. figs. 4, 5. 

1890. Gonatodus punctatus, R. H. Traquair, Proc. Roy. Soc. Edinb., vol. xvii. p. 391. 

1891. Gonatodus punctatus, A. S. Woodward, Cat. Foss. Fishes Brit. Mus., pt. ii. p. 434. 

When the preliminary note on these fishes was written in 1904, only one relic of 
this beautiful Palseoniscid had been found at Gullane, namely, the dentary element of 
the right ramus of the mandible, showing the characteristic dentition, but subsequently 
four additional examples were obtained by Mr Macconochie. These are — two examples 
of the entire fish ; another showing the scales of the anterior part of the body in 
apposition ; while the fourth is the impression of the two rami of the mandible with the 
branchiostegal rays. Of the two specimens of the entire fish, one (PL II. fig. 1) is 
especially interesting as being the smallest example of the species which I have ever 
seen, its length being a little over 3 inches and its greatest depth slightly over 
1 inch. The other entire specimen is larger, but somewhat distorted in shape. As 
regards the one showing the flank scales in apposition, with the operculum and some 
elements of the shoulder girdle, the scales are remarkably smooth and free from 
striations, but there can be no doubt as to the correctness of the specific determination. 

Genus Elonichthys, Giebel. 

Elonichthys Robiso7\i, Hibbert, sp. PL I. fig. 1. 

1835. Palaioniscus Robisoni, Hibbert, Trans. Roy. Soc. Edinb., vol. xiii. p. 191, pi. vi. figs. 6, 7, 

pi. vii. figs. 1-3. 

1835. Palceoniscus Robisoni, Agassiz, Poiss. Foss., vol. ii. pt. i. p. 88, pi. x.a figs. 1, 2. 

1835. Palaioniscus striolatus, Agassiz, ibid., p. 91, pi. x.a figs. 3, 4. 

1835. Amblypterus nemopterus, Agassiz, ibid., p. 107, pi. iv.b figs. 1, 2. 

1835. Amblypterus punctatus, Agassiz (pars), ibid., p. 109, pi. iv.c figs. 3, 5-8. 

1835. Pygo/derus Bucklandi, Agassiz in S. Hibbert, op. cit., p. 217, pi. vii. fig. 2. 

1814. Pygopterus Bucklandi, Agassiz, Poiss. Foss., vol. ii. pt. ii. p. 77. 

1877. Elonichthys Robisoni, R. H. Traquair, Qu. Journ. Geol. Soc, vol. xxxiii. p. 553. 

1877. Elonichthys intermedins, R. H. Traquair, Proc. Roy. Soc. Edinb., vol. ix. p. 279. 

1877. Elonichthys ovatus, R. H. Traquair, ibid., p. 427. 

1877. Elonichthys nemopterus, R. H. Traquair, ibid., p. 278, and Qu. Journ. Geol. Soc, vol. xxxiii. p. 553. 


1877. Elonichthys striolatus, R. H. Traquair, ibid., p. 553, and Ganoid Fishes Brit. Carb. Form. (Pal. 

Soc), p. 57, pi. vii. figs. 4-15. 
1880. Elonichthys tenuiseratus, R. H. Traquair, Proc. Boy. Phys. Soc. Edinb., p. 119. 
1880. Elonichthys Dunsii, R. H. Traquair, ibid., p. 126. 

1891. Elonichthys Robisoni, A. S. Woodward, Cat. Foss. Fishes Brit. Mus., pt. ii. p. 495. 
1891. Elonichthys Bucldandi, A. S. Woodward, ibid., p. 497. 
1901. Elonichthys Bobisoni, R. H. Traquair, Gan. Fishes Brit. Carb. Form., pp. 62-70, — including as 

variations, striolotus, Bucklandi, intermedins, nemopterus, tenui striatus, and Dunsii. 

A considerable number of more or less fragmentary specimens are referable to this 
species, so common in the estuarine beds of the Lower Carboniferous series in Scotland. 
Among these, the nearest approach to a complete fish is represented in PI. I. fig. 1. 
This example appears to be referable to the variety nemopterus (see my Pal. Soc. 
monograph, 1901, p. 68), characterised by its delicately marked scales and slender 
fin-rays with distant articulations, but most of the others approach rather more the type 
of intermedins (op. at., p. 67). But the varieties of E. Robisoni all seem to pass 
insensibly into each other, as is natural in so common a species. 

Elonichthys striatus, Agassiz, sp. 

1835. Amblypterus striatus, Agassiz, Poiss. Foss., vol. ii. pt. i. p. iii., pi. iv.b figs. 3-6. 
1857. Bhabdolepis striatus, Troschel, Verb, natnrh. Verein. preuss., Rheinl. und Westfal., p. 16. 
1877. Gosmoptychius striatus, R. H. Traquair, Qu. Jour. Geol. Soc, vol. xxxiii. p. 553 ; and Gan. 
Fishes Brit. Carb Form. (Pal. Soc), p. 43, pi. ii. fig. 7, pi. iii. figs. 1-8. 

1890. Elonichthys striatus, R. H. Traquair, Proc. Boy. Soc. Edinh., vol. xvii. pp. 390, 396. 

1891. Elonichthys striatus, A. S. Woodward, Cat. Foss. Fishes Brit. Mtis., pt. ii. p. 491. 

This is perhaps the commonest species of fish in this locality, the Survey collection 
containing a great many specimens, some of which are nearly entire. Most of 
these are from bed a ; but its presence in bed b has been proved not only by Mr 
Macconochie, but also by Messrs Bowie and W. T. Gordon. 

Elonichthys striatulus, Traquair. PL I. figs. 2-5. 
1904. Elonichthys striatidus, R. H. Traquair, in Summ. Prog. Geol. Survey for 1903, p. 121. 

The specimen represented in PL I. fig. 2, the only one which has hitherto occurred 
at Gullane, measures a little over 2 inches as it stands, but its original length may be 
estimated at 2^,* as the tip of the caudal extremity and also of the snout are both 
cut off by the edges of the stone. The length of the head would probably be to 
the total as one in four ; its roof-bones are ornamented by comparatively coarse and 
distant ridges, with some tubercles on the parietal region ; the external sculpture seen 
on the facial bones and on the elements of the shoulder girdle consists also of ridges of 

* In the diagnosis of this species given in the Summary of Progress a strange misprint occurs, " 21 " inches 
being given as the estimated length of the specimen instead of " 2j." 


a similar character to those on the cranial roof. The scales of the flank (fig. 3) are 
ornamented with fine yet well-marked ridges, which pass across the exposed surface from 
front to back in a direction mainly parallel with the upper and lower margins ; in 
many cases the ridge is double, the two divisions uniting close to the posterior margin 
of the scale ; in the caudal region (fig. 4) the ridges become more oblique in their 
direction. Several large azygous scales occur in front of the dorsal fin. This fin, 
situated about the middle of the back, is proportionally large, its rays are very slender, 
and with unusually distant articulations ; the joints are marked occasionally with a 
very delicate longitudinal furrow, or it may be with two ; the form of the fin would be 
triangular acuminate were not the apical portion largely cut off by the edge of the stone. 
The anal fin is lost ; the caudal is heterocercal, inequilobate, deeply cleft, and with 
distantly articulated slender rays similar to those of the dorsal ; fin-fulcra few and very 
oblique. Paired fins not preserved. 

On looking over some undetermined Eskdale material in the collection of the Royal 
Scottish Museum, I found a small fish from Glencartholm which is certainly specifically 
identifiable with the above-described example from Gullane, and which is represented in 
PI. I. fig. 5. It is slightly larger, as it measures nearly 3 inches in length, and is 
more perfect, inasmuch as its contour is seen from the tip of the snout to the termina- 
tion of the upper lobe of the tail-fin. The length of the head is contained about four 
times in the total ; its superficial bones are sculptured, as in the Gullane specimen, with 
comparatively coarse and distant ridges, and the form and ornament of the scales is also 
the same. As in the above-described specimen, the paired fins are not shown, while, in 
addition, the dorsal is wanting ; on the other hand, the anal is present and the caudal is 
complete. We have again the same character of the fin-rays, — as before, they are very 
slender and very distantly articulated, while the fulcra in front of the anal and caudal 
are few, elongated, and oblique. 

I have placed this little fish in the genus Elonichthys on account of its general 
aspect and the form and position of the unpaired fins, though the condition of the fin- 
fulcra deviates considerably from that which is usual in the genus. In all its details it 
is strikingly distinct from every other known species. 

The specimen, PI. II. fig. 2, collected by Mr Macconochie at Gullane, is from 
bed a. 

Genus Rhadinichthys, Traquair. 
Rhadinichthys ornatissimus, Agassiz, sp. 

1835. Palxoniscus ornatissimtis, Agassiz, Poiss. Foss., vol. ii. pt. i. p. 92, pi. x.a figs. 6, 8 (non figs. 5, 7). 
1877. Rhadinichthys ornatissimus, R. H. Traquair, Qu. Journ. Geol. Soc, vol. xxxiii. p. 559, and 

Proc. Roy. Soc. Edinb., vol. ix. p. 432. 
1877. Rhadinichthys lepturus, R. H. Traquair, Proc. Roy. Soc. Edinb., vol. ix. p. 437. 

1890. Rhadinichthys ornatissimus, R. H. Traquair, Proc. Boy. Soc. Edinb., vol. xvii. pp. 391, 397. 

1891. Rhadinicldhys ornatissimus, A. S. Woodward, Cat. Foss. Fishes Brit. Mus., pt. ii. p. 462. 


One fragmentary specimen has occurred, showing the head with a portion of the 
front of the body, the scales of which are considerably mashed up together. However, 
what can be seen of the ornamentation of these scales, along with the impressions of 
the bones of the head, shows conclusively with what species we have to deal. 

Collected by Mr Macconochie in bed a. 

Rhadinichthys brevis, Traquair. PI. II. figs. 3-5. 

1877. Rhadinichthys brevis, R. H. Traquair, Proc. Roy. Soc. Edinb., vol. ix. p. 440. 
1891. Rhadinichthys brevis, A. S. Woodward, Gat. Foss. Fishes Brit. Mm'., pt. ii. p. 462. 
1901. Rhadinichthys brevis, R. H. Traquair, Geol. May. (4), vol. viii. p. 111. 

In the remarks appended to my original description of the species from the beds at 
Wardie, I compared it specially with Rh. Geikiei, Traq., from Colinton. But as I have 
since then adopted the view that Rh. Geikiei represents but a young stage of Rh. 
carinatus, Agassiz, it is with the adult of the last-named species that the comparison 
must be carried out. Compared, then, with R. carinatus, Rh. brevis is, as the name 
implies, proportionally shorter, the unpaired fins are larger, the distance between the 
commencement of the anal and that of the caudal is proportionally smaller, the ridges 
on the cranial roof-bones are coarser, though in this latter respect there is some 
variation in Rh. brevis. It is undoubtedly a good species, and occurs not only at the 
original locality, Wardie, and here, at Gullane, but also in Fifeshire (Geol. Mag., t. cit. 
ant., p. 111). 

Collected by Mr Macconochie from bed a ; — a slightly doubtful specimen has also 
been obtained by Mr W. T. Gordon in bed b. 

Rhadinichthys elegantulus, Traquair. 

1881. Rhadinichthys Geikiei, R. H. Traquair, Trans. Roy. Soc. Edinb., vol. xxx. p. 25, pi. i. figs. 13-18. 
1881. Rhadinichthys Geikiei, var. elegantulus, R. H. Traquair, ibid., p. 27, pi. ii. figs. 1-5. 
1881. Rhadinichthys delicatulus, R. H. Traquair, ibid., p. 29, pi ii. figs. 6-9. 

1890. Rhadinichthys elegantulus (with var. delicatulus), R. H, Traquair, Proc. Roy. Soc. Edinb., 

vol. xvii. p. 398. 

1891. Rhadinichthys elegantulus. A. S. Woodward, Cat. Foss. Fishes Brit. Mus., pt. ii. p. 464. 

To this species I refer a small and not quite perfect specimen collected by 
Mr Macconochie in bed a. I certainly cannot distinguish it from the Glencartholm 
fish whose synonymy is given above. 

Rhadinichthys formosus, Traquair. PL I. fig. 6, PI. II. figs. 6-8. 

1904. Rhadinichthys formosus, R. H. Traquair, in Summary of Progress Geol. Survey for 1903, p. 122. 

The original specimen of this form is represented in PI. II. fig. 6 ; unfortunately 
it is not perfect, as nearly the whole of the upper and a great portion of the lower lobe 
of the caudal fin is lost. As it is, the length is 3f inches ; but when entire the fish 


must have measured 4^ inches. The length of the head is contained three and a half 
times in the distance from the tip of the snout to the commencement of the lower lobe 
of the caudal fin, and is equal to the greater depth of the body between the dorsal fin 
and the occiput. The cranial roof-bones are marked with fine contorted and interrupted 
strise, and a striated ornament characterises also the facial bones, namely, the maxilla, 
mandible, and branchiostegal rays. The dorsal fin is situated behind the arch of the 
back, and more than twice the length of the head from the end of the snout ; the anal 
commences a little further back ; both fins are large, triangular-acuminate, with many 
slender rays which are distantly articulated, the joints being often marked with a 
delicate furrow close to, and parallel with, the posterior margin. The fulcra in both 
fins are very well marked. The caudal is, as already indicated, broken off near its 
commencement ; indications both of pectoral and ventral rays are seen at their proper 
places, but the fins themselves cannot be described. The scales are moderate in size, 
those on the flank showing an ornament which to some extent resembles that of the 
scales of Elonichthys pulcherrimus from the Glencartholm beds, the exposed surface 
being covered with closely-set fine ridges, those on the posterior-superior part passing 
obliquely downwards and backwards, while on the anterior-inferior aspect they pass 
nearly horizontally backwards, and parallel to the inferior margin. On the hinder 
border of each scale these ridges end in delicate denticulations. As regards the scales 
posterior to the region of the ventral fins, the ornament tends to become less pronounced, 
and reduced to a set of grooves passing obliquely across the surface from front to back, 
but the scales along the origin of the anal fin, and also of the dorsal, as seen in 
the counterpart, are strongly ridged. All over the body the scales are singularly 
undisturbed, so that the proportions of the fish are accurately shown. 

In PI. I. fig. 6 is represented another specimen, which, according to its scale 
ornamentation and other details, I identify specifically with the one described above. 
Its proportions seem, however, at first sight, rather different, the dorsal fin appearing 
relatively further forward, the caudal disproportionately large, while the depth of the 
body at the shoulder is considerably greater than the apparent length of the head. 
But, on a careful examination of the specimen, it is evident that these appearances are 
due to distortion by " shortening up," the scales of the region behind the head being- 
confused, so that the body at this place is, as it were, " telescoped." The specimen 
measures 4|- inches in length, but I should add on at least one more inch to express its 
original dimensions, so that we have now before us a larger example than the previous 
one. I figure this specimen on account of the perfect caudal extremity which it 
displays. The fin is deeply cleft and inequilobate, the prolongation of the body axis 
along the upper lobe being proportionately stronger than in the type species of the 
genus (Rh. ornatissimus). 

Still another specimen exists in the Survey Collection, but less perfect than either 
of the two examples described above. It is also shortened up in a manner somewhat 
similar to that noticed in the second one. 


I have placed this species in Rhadinichthys owing to its resemblance in external 
form to Rh. brevis, although the condition of the pectoral fin cannot be ascertained. 
From bed a, collected by Mr Macconochie, and also by myself. 

Nematoptychius Greenocki, Agassiz sp. 

1844. Pygopterus Greenockii, Agassiz, I'oiss. Foss., vol. ii. pt. ii. p. 78 (undefined). 

1867. Pygopterus Greenockii, R. H. Traquair, Trans. Roy. Soc. Edirib., vol. xxiv. p. 701, pi. xlv, 

figs. 1-7. 
1872. Pygopterus elegans, C. W. Peach, Rep. Brit. Assoc, 1871, Trans. Sect., p. 109 (name only). 
1875. Nematoptychius Greenockii, R. H. Traquair, Ann. Mag. Nat. Hist. (4), vol. xv. p. 258, pi. xvi. 

figs. 7-11. 
1877. Nematopty chius gracilis, R. H. Traquair, Proc. Roy. Soc. Edirib., vol. ix. p. 262 ( = young form of 

N. Greenockii). 
1877. Nematopty chius Greenockii, R. H. Traquair, Qu. Journ. Geol. Soc, vol. xxxiii. p. 577, and 

Ganoid Fishes Brit. Carb. Form. (Pal. Soc), pi. i. figs. 7-11. 

1890. Nematoptychius Greenocki, R. H. Traquair, Proc. Roy Soc. Edinb., vol. xvii. pp. 391, 398. 

1891. Nematoptychius Greenocki, A. S. Woodward, Gat. Foss. Fishes Brit. Mus., pt. ii. p. 458. 

Fragmentary specimens of this interesting Palseoniscid, so extensively distributed in 
the Lower Carboniferous rocks of the East of Scotland, have been found in Bed b, both 
by Mr Macconochie and by Mr W. T. Gordon. 

Family Platysomid^e. 

Genus Wardichthys, Traquair. 

Wardichthys cyclosoma, Traquair. PI. II. fig. 9. 

1875. Wardichthys cyclosoma, R. H. Traquair, Ann. Mag. Nat. Hist. (4), vol. xv. p. 262, pi. xvi. 

figs. 1-5. 
1879. Wardichthys cyclosoma, R. H. Traquair, Trans. Roy. Soc. Edinb., vol. xxix. p. 361, pi. iv. 

figs. 12-15. 
1881. Wardichthys cyclosoma, T. Stock, Ann. Mag. Nat. Hist. (5), vol. vii. p. 490. 
1891. Wardichthys cyclosoma, A. S. Woodward, Gat. Foss. Fishes Brit. Mus., pt. ii. p. 535. 

The genus Wardichthys, with the single species cyclosoma, was instituted by me 
in 1875 for a small Platysomid fish from the Lower Carboniferous shales, exposed at 
Wardie beach, near Edinburgh. Only a single specimen was then known, and of this 
I gave a minute description in the Annals and Magazine of Natural History for April 
of the year named. Again, in my Essay on the " Structure and Affinities of the 
Platysomidse," published in the Transactions of the Royal Society of Edinburgh for 
1879, I recapitulated the principal characters of the genus, and showed that of all the 
generic forms of Platysomidae it was most nearly allied to Mesolepis. 




The original specimen showed the general shape of the body and the character of 
the squamation, some of the osteological details of the head, the dorsal fin, and a few 
traces of the anal. Another specimen, subsequently obtained by the late Mr T. Stock, 
and of which the counterpart is now in the Royal Scottish Museum, shows distinctly 
the pectoral and ventral fins, but only traces of the anal, while the caudal is again 
wholly wanting, having been cut off by the edge of the ironstone nodule in which the 
fish is contained. 

But quite a number of specimens of Wardichthys cyclosoma have turned up at 
Gullane, and in several of these the caudal as well as the anal fin is well shown, so that 
now the material has been obtained on which to construct the restored figure, which I 
herewith append. The generic diagnosis may also be amended as follows : — 

Fig. 1. — Restored outline of Wardichthys cyclosoma, Traquair. pa., parietal; /., frontal; sq., squamosal; pt.f., post- 
frontal; op., operculum; s.op., suboperculum ; mx., maxilla; p.t., post-temporal;, supraclavicular; cl., clavicle; 
p., pectoral ; v., ventral ; d., dorsal ; a., anal ; c, caudal. 

Genus Wardichthys, Traquair. Body deep, nearly circular in profile ; ventral fins 
present, situated nearly exactly between the pectorals and the anal. Dorsal fin small, 
arising considerably behind the top of the rounded arch of the back, and terminating 
near the commencement of the tail pedicle ; its base somewhat extended in proportion 
to its height. Anal, placed a little further back than the dorsal, short based, triangular- 
acuminate in form. Caudal comparatively small, heterocercal, deeply cleft. All the 
fins with distinct fulcra ; their rays closely set and closely articulated. Cranial 
osteology allied to that of Mesolepis, but the dentition is still unknown. Scales of 
the middle of the flank high and narrow, more equilateral behind and towards the 
dorsal and ventral margins ; covered area broad, exposed surface ornamented by a close 
tuberculation, the tubercles often coalescing into short ridges ; azygous scales in front 
of the dorsal and anal fins, with well-marked backwardly-directed spur-shaped points. 
Only species, W. cyclosoma, Traq. 


The new points here brought out are — the presence of pectorals and ventrals, and 
the contour of the anal and caudal fins. Though the dentition is still unknown, the 
affinity of the genus to Mesolepis is obvious, as already remarked in my paper on the 
Platysomidse, but the small size of the dorsal fin and its backward position are marks 
of sufficient importance to justify the validity of Wardichthys. 

For purposes of comparison I add a restored outline of Mesolepis, of which only the 
head part has been previously published in my paper on the Structure and Affinities of 
the Platysomidse. 

All the specimens from the Gullane locality in the Survey Collection are from 
bed a, and represent fishes from 2 to 4 inches in length. From bed c (the " crustacean " 
bed) Mr F. H. Butler, of London, obtained a nearly perfect specimen, now in 

Fig. 2.— Restored outline of Mesolepis scalaris, Young, from the North Staffordshire Coalfield, e., dermal ethmoid - 
»., nasal opening; p./., pre-frontal ; p.op., preoperculum ; s.o., suborbital; mn., mandible;, premaxilla. Other 
lettering as in fig. 1. 

the British Museum. I am indebted to Dr A. S. Woodward for the privilege of 
examining this specimen, the largest which I have seen, as it measures a fraction over 
5 inches in length. 


Having now gone over, from a zoological standpoint, the fishes collected at this 
interesting locality, it remains for me to say a few words regarding their interest to the 

It is, of course, for the officers of the Geological Survey to settle the horizon of 
these fish-bearing beds by correlating the various lights thrown on them by stratigraphy 
as well as by palaeontology, and here it must also be remembered that they contain 
other fossils besides fishes. Nevertheless, it may be permitted to me to point out the 



result of a comparison of this fish-fauna with that of other localities in Scotland in 
which estuarine fishes of Lower Carboniferous age occur. 

First of all, we may give a complete list of the fishes which have hitherto been 
found at this locality, indicating also the respective beds in which they have 

Bed a. 

Bed b. 


Acanthodes, sp., ..... 



Rhizodus Hibberti, 



Coelacanthus, sp., . 




Gonatodus punctatus, 




Elonichthys Robisoni, 




„ striatus, 




„ striatidus, 




RhadiniclUltys ornatissimus, 



„ brevis, 




„ elegantulus, 




„ formosus, 




Nematoptychius Greenoclri, 




Wardiehthys cyclosoma, 



Here I may first draw attention to the fact that the list contains no representative 
of the Elasmobranchii, with the exception of the peculiar group of Acanthodei. 

Theij I need hardly point out that the assemblage is one of Lower Carboniferous 
estuarine forms. Not one of the named species can be noted as indicating an Upper 
Carboniferous horizon. 

Among those named species we are struck with the presence of one rare form — 
Wardiehthys cyclosoma — hitherto found with certainty only in one locality, and that 
locality is Wardie, near Granton, from, which a division of the Calciferous Sandstone 
series of the Edinburgh district, namely, the Wardie Shales, takes its name. 

In my paper " On the Distribution of Fossil Fish-remains in the Carboniferous 
Rocks of the Edinburgh District," * I have given a list of the fossil fishes of Wardie 
beach, which I reproduce here : — 

Pleuracanthus, sp. 
Acanthodes sulcatus. 
Megalicldhys, sp. 
Rhizodus Hibberti. 
Gonatodus punctatus. 
Elonichthys Robisoni. 

Rh adinichthys ornatissimus. 

,, carinatus. 

,, brevis. 

Nematoptychius Greenocki. 
Eurynotus crenatus. 
Wardiehthys cyclosoma. 

Here we find that, in this list of fourteen species, eight out of the thirteen from 

'• Trans. Roy. Soc. Edinb., vol. xl., pt. iii., No. 28, p. 690. 



Gullane are repeated, namely, Rhizodus Hibberti, Gonatodus punctatus, Elonichthys 
Robisoni, E. striatus, Rhadinichthys ornatissimus, Rh. brevis, Nematoptychius 
Greenocki, and Wardichthys cyclosoma. And of these eight species, four, namely, 
Gonatodus punctatus, Elonichthys Robisoni, Elonichthys striatus, and Nematoptychius 
Greenocki, are the commonest of all the fishes which are to be found in the Wardie 

I now give a second table showing the proportion of the specifically determined 
Gullane fishes represented in Lower Carboniferous estuarine rocks of the Edinburgh 
•district belonging to higher horizons than the Wardie Shales. 


<£> Ph 

2 di 

Named Species from Gullane. 

.8 «5 

1 c 
J -S 




St A 


& a 

h-J en 

Edge Coal 

•H o 


nk o 

Rhizodus Hibbeiii, 






Gonatodus punctatus, 






Elonichthys Robisoni, 







„ striatus, 







,, striatulus, . 







Rhadinichth i/s ornatissimus, 







„ brevis, 







„ elegantulus, 







„ formosus, 







Nematoptych ius Greenocki, 







Wardichthys cyclosoma, 











From this table, that of the eleven named species found in the Gullane beds, — 

Eight occur in the Wardie Shales. 

Five in the Burdiehouse Limestone. 

Four in the Oil Shales between the Burdiehouse Limestone and the Lower Marine 
Limestone series. 

Three in the Lower Limestone series. 

Four in the Edge Coal series. 

Two in the Upper Limestone series. 

Or, to express the matter in different words, while nearly four-fifths of the named 
Gullane species are represented at Wardie, these species become fewer and fewer as we 
proceed upwards in the Lower Carboniferous of the district, till in the Upper Limestone 
series only two, or less than one-fifth, has as yet been met with. 

In my paper already quoted, I have remarked (p. 698) "that it is scarcely possible 
to mark out satisfactory zones by means of the fishes, so far as the Lower Carboniferous 
rocks are concerned " ; but I have also qualified this statement by another, namely, 
that " some amount of difference there is between the lists from the top and from the 


bottom of the fish-bearing series" (p. 701). And I have shown that while some species 
like Nematoptychius Greenocki and Eurynotus crenatus persist from bottom to top of 
these estuarine fish-bearing rocks, some have as yet only occurred in the lower, others 
in the upper, horizons as usually adopted. 

If, then, fish-remains are to count in the determination of the horizon of the Gullane 
beds, I would point out that their fish-fauna show the greatest amount of agreement 
with that of the Wardie Shales of all the horizons in the Lower Carboniferous of 
Central Scotland in which fishes have been found. 

Finally, we have seen that out of eleven named Gullane species, eight occur at 
Wardie. What of the other three, which have not been found in any other locality near 
Edinburgh ? 

Of these, one is peculiar to the East Lothian locality, namely, Rhadinichthys 

The remaining two, namely, Elonichthys striatulus and Rhadinichthys elegantulus, 
form part of the peculiar Lower Carboniferous fish-fauna of Glencartholm, on the Eiver 
Esk, near Langholm. Two explanations of the peculiarity of this fish-fauna have been 
suggested, — either the existence of a land barrier, or a difference in horizon, these 
Eskdale beds being supposed to be of older date than the Granton and Craigleith 
Sandstones. The latter explanation does not seem probable in the face of the fact 
that such well-known Central Scottish species as Or acanthus armigerus, Traq., and 
Tristychius minor, Port!., have also been found at Glencartholm, while the determina- 
tion of two further species as common to the last-named locality and Gullane is further 
evidence against the idea, so far as fishes are concerned. If, however, the existence of 
a land barrier was the cause of the phenomenon in question, it is clear that it was not 
sufficient to prevent the intermixture of species to a certain extent. The Glencartholm 
beds were certainly not deposited in a land-locked lake at all events, as the occurrence 
of numerous marine shells in them amply testifies. The further development of the 
subject I must leave to the Officers of the Geological Survey. 


[All the figures are taken from specimens from Bed a, Gullane, except fig. 5 on Plate I., which 
illustrates a specimen from Glencartholm, near Langholm, Eskdale.] 

Plate I. 

Fig. 1. Elonichthys Rohisoni (Hihhert). Natural size. Gullane. (Geol. Surv. Scot., M 293 f.) 

Fig. 2. Elonichthys stria/ utus, Traq. Natural size. Same locality. (Geol. Surv. Scot., M 4409 E.) 

Fig. 3. Flank scales of the above specimen, magnified 12 diameters. 

Fig. 4. Caudal scale from the same, magnified 12 diameters. 

Fig. 5. Elonichthys striatulus. Traq. Natural size. From Glencartholm, Eskdale, in the Royal 
Scottish Museum. 

Fig. 6. Rhadinichthys formosus, Traq., distorted by "shortening up" of the anterior part of the body. 
Natural size. Gullane. (Geol. Surv. Scot., M 292 f.) 


Plate II. 

Fig. 1. Gonatodus punctatus (Agassiz). Young specimen; natural size. Gullane. (Geol. Surv. Scot., 
M 297 f.) 

Fig. 2. Inner surface of the dentary element of the mandible of an adult specimen of Gonatodus 
punctatus, magnified 2 diameters. Some scales are scattered over the proximal extremity of the bone, and 
the teeth are considerably obscured by matrix. Same locality. (Geol. Surv. Scot., M 4437 E.) 

Fig. 3. Rhadinichthys brevis, Traq. Natural size. Same locality. {Geol. Surv. Scot., M 4411 E.) 

Fig. 4. Another specimen of Rh. brevis. Natural size. Same locality. (Geol. Surv. Scot., M 296 f.) 

Fig. 5. Flank scales of Rh. brevis from another specimen, magnified 8 diameters. Same locality. 

Fig. 6. Rhadinichthys formosus, Traq. Natural size. Same locality. (Geol. Surv. Scot., M 4418 E.) 

Fig. 7. Scale from the flank of the same specimen, magnified 6 diameters. 

Fig. 8. Scale from the caudal region of the same specimen, magnified 6 diameters. 

Fig. 9. Wardichthys cyclosoma, Traq. Natural size. Same locality. (Geol. Surv. Scot., M 4384 E.) 

■Roy. Soc.Edin^ 


ATE r. 

Vol. XL V] 


cJ. Gr 

in lap. 




■00 HX) 

( 119 ) 

V. — The Marine Mollusca of the Scottish National Antarctic Expedition. By James 
Cosmo Melvill, M.A., F.L.S., F.Z.S., and Robert Standen, Assistant Keeper, 
Manchester Museum. Communicated by Professor J. C. Ewart, F.R.S. (With 
a Plate.) 

(MS. received July 5, 1907. Read July 15, 1907. Issued separately December 21, 1907.) 

The marine Mollusca obtained during the Scottish National Antarctic Expedition of 
s.s. Scotia are particularly interesting, not merely on account of certain benthal forms 
having been for the first time brought to light from unusual depths, but also as 
demonstrating an extended geographical range for certain meridional species already 
known ; while the occurrence of sundry British and North European individuals in the 
Falkland Islands — perhaps brought there adventitiously — is, to some extent, a new 
feature in the fauna. Regarding the deep-sea species just alluded to, the dredging of 
a second example of the hitherto unique Guivillea alabastrina, Watson, from a station 
not far from the original Challenge)- locality, is very noteworthy ; while we may at the 
same time signalise a Columbarium, Cuspidaria, Chrysodomus, and Dentalium, from 
1775, 2645, 1775, and 1410 fathoms respectively. 

The attention of the scientific world has for some years become more and more 
centred upon the Arctic and Antarctic polar regions, and, of them, the latter till lately 
presented the most virgin field for research, both physical, geographical, and biological. 
Every year fresh expeditions seem to be planned, equipped, and launched for the 
purpose of critical investigation and exploration, with most beneficial results ; and the 
day may not now be far distant when it will be possible to draw up monographs of the 
various zoological groups of the Antarctic area, including, we hope, one of the Mollusca. 
In the meantime, we give below a collated bibliographical resume of the majority of the 
treatises published on the subject during the past thirty years. 

In conclusion, we would express our obligations primarily to Mr Edgar A. Smith, 
I.S.O., of the British Museum (Nat. Hist.), who has, with us, examined every critical 
species in the collection. To Mr Ernest R. Sykes, F.L.S., and Dr W. E. Hoyle, F.R.S.E., 
we are also indebted for the loan of various books and pamphlets bearing upon the 
subject. To Dr Georg Pfeffer and Dr Hermann Strebel, of Hamburg, our thanks 
are due for reporting on several forms submitted to them, two of which, obtained 
likewise by the Swedish South Polar Expedition, and differentiated, but not yet 
published, by Dr Strebel, we have been asked by him to describe at the present 

And last, but by no means least, we are most grateful to Mr W. S. Bruce, F.R.S.E., 
for having entrusted his Molluscan collections to us, and for much kind assistance in the 
way of maps and general information. 



A. Regio Antarctica — including Gough Island. 


Order Amphineura. 

Sub-order polyplacophora. 

Callochiton illuminatw (Reeve) 
Tonicia atrata (Sowb.). 

Order Prosobranchiata. 

Sub-order diotocardia. 

Family Acmseidse. 
Acmsea ceciliana, Orbigny. 

Family Patellidae. 
Patella senea, Marty n, var. deaurata, Gmel. 
,, fuec/iensis, Reeve. 
,, polaris, Hombr. and Jacq. 

Section Zygobranchiata. 

Family Fissurellidse. 
Fissurella oriens, Sow. 
,, picta, Gmel. 
Tugalia antarctica, sp. n. 

Section Azygobranchiata. 

Family Trochidse. 
Photinula expansa (Sow.). 
,, tseniata (Wood). 
,, violacea (King). 
Valvatella antarctica (E. Lamy). 

Sub-order monotocardia. 

Section (a) Ptenoglossa. 

Family lanthinidse. 
lanthina exigua, Lamarck. 


Section (b) Taenioglossa. 

Family Naticid&. 
Natica (Lunatia), sp. 

Family Capulidse. 

Calyptvsea costellata, Phil. 
Crepidula dilatata, Lamk. 

Family Littorinidse. 

Littorina (Lsevilitorina) caliginosa (Gould). 

,, ,, coriacea, sp. n. 

„ (Pellilitorina) pellita, v. Marts. 

,, ,, setosa (Smith). 

Lacuna divaricata, Fabr. 
,, notorcadensis, sp. n. 

Family Rissoidae. 

Rissoa adarensis, Smith. 

(Cingula) cingillus (Mont.). 

edgariana, sp. n. 

fraudulenta, Smith. 

parva (Da Costa). 

(Onoba) scotiana, sp. n. 

(Manzonia) zetlandica (Mont.). 
Eatonlella kerguelensis, Smith. 

Family Litiopidzs. 
Litiopa melanostoma, Kang. 

Family Ceriihidse. 

Cerithium georgianum, Pfeft'er. 

„ pullum, Phil. 
Cerithiopsis malvinarum (Strebel, MS.), sp. n. 

Family Tritonidw. 
Gyrineum vexillum (Sow.). 

Section (c) Gymnoglossa. 



Section (d) Rachiglossa. 

Family Muricidas. 

Trophon hrucei, Streb. 

cinguliferus, Pfeffer. 

crispus (Couth.). 

geversianus (Pallas). 

hoylei, Streb. 

liratus (Couth.). 

minutus (Strebel, MS.), sp. n. 

philippianus, Dunker. 

Family Nassidse. 
Nassa (Ilyanassa) vallentini, sp. n. 

Family Buccinidee. 

Chrysodomus (Sipho) archibenthalis, sp. n. 
,, ,, crassicostatus, sp. n. 

Neobuccinum eatoni (Smith). 
Euthria fuscata (Brug.). 

,, magellanica (Philippi). 

,, michaelseni, Streb. 

Family Volutidse. 
Voluta (Cymbiola) ancilla (Sol.), 
Guivillea alabastrina, Watson. 

Section (e) Toxoglossa. 

Family Cunidas. 

Columbarium benthocallis, sp. n. 
Mangilia costata (Donovan). 

Family Cancellariidai. 
Admete magellanica, Strebel. 

Order Opisthobranchiata. 
Sub-order tectibranchiata. 

Family Tornatinidse. 
Retusa truncatula, (Brug.). 


Section Siphonarioidea. 

Family Siphonariidae. 
Siphonaria redimiculum, Reeve. 

Order Pulmonata. 

Sub-order basommatophora. 

Familv Auriculidse. 
Marinula nigra, Phil. 


Dentalium eupatrides, sp. n. 
,, shoplandi, Jouss. 

Order Protobranchiata. 

Family Nuculidse. 

Nucula minuscula, Pfeffer. 
Yoldia eightsi (Couth.). 

Order Filibranchiata. 

Sub-order anomiacea. 

Family Anomiidse. 
Anomia ephippium, Linn. 

Sub-order arcacea. 

Family Arcadse. 

Area (Bathyarca) strebeli, sp. n. 
Lissarca notorcadensis, sp. n. 
,, rubrqfusca, Smith. 

Sub-order mytilacea. 

Family Mytilidse. 

Mytilus edulis, Linn. 

,, ma.gellanicus, Chemnitz. 
ovalis. Lamarck. 


Philobrya meridionalis (Smith). 

,, quadra ta (PfefYer). 

,, wandelensis, E. Lamy. 

Modiolarca mesembrina, sp. n. 

Order Pseudolamellibranchiata. 

Family Pectenidze. 

Pec ten colbecki, Smith. 

,, multicolor, sp. n. 

,, 1 patagonicus, King. 

„ pteriola, sp. n. 
Amussium octodecim-liratum, sp. n. 

Family Limidx. 

Lima (Mantellum) goughensis, sp. n. 
,, (Limatula) pygmxa, Philippi. 

Order Eulamellibranchiata. 

Sub-order submytilacea. 

Family Lucinidse. 
Cryptodon falklandicus, Smith. 
Cyamium antarcticum, Philippi. 

,, falklandicus, Mel v. and St. 

Family Erycinidse. 
Lassea consanguinea (Smith). 
Kellia cycladiformis, Desh. 

,, lamyi, nom. nov. 
? Scacchia plenilunium, sp. n. 

Sub-order tellinacea. 

Family Tellinidse. 
Tellina (Maera) pusilla (Philippi). 

Sub-order veneracea. 

Family Veneridse. 
Chione philomela (Smith). 
Tapes (Amygdala) exalbida (Chem.). 


Sub-order myacea. 
Family Glycimeridse. 
Saxicava arctica (L.), var. antarctica, Phil. 

Sub-order anatinacea. 
Family Lyonsiidse. 
Lyonsia cuneata (Gray). 

Family Anatinidze. 
Anatina elliptica, King and Brod. 

Order Septibranchiata. 
Family Cuspidariidw. 
Cuspidaria brucei, sp. n. 

B. From Ascension Island. 

Lotorium grandivnaculatum (Reeve). 
Nerita (Thelicostyla) ascensionis, G-mel. 

C. From Saldanha Bay, South Africa. 
Chzetopleura papilio (Spengler). 
Ischnochiton tigrinus (Krauss). 
Chiton nigrovirens, Blainville. 
Acanthochites garnoti, Blainville. 
Bullia annidata (Lam.). 

,, laevigata, Chem. ( = lwoissima, Gmel.). 
Mytilus edulis, L., var. meridionalis, Krauss. 
Mytilicardia {lliecalia) concamerata, Brug. 
Tapes pullastra (Montagu). 
Oxy stele impervia, Minhe. 

,, tigrina, Chem. 
Crepidida hepatica, Desh. 
Gyrineum (Argobuccinum) argus (Gmel.). 
Phos plicosus, Dunker. 
Cominella limbosa (Lam.). 
Patella compressa, L., var. miniata, Born. 

,, (Scutellastra) granatina, L. 

„ ,, oculus, Born. 

Helcion pectinatum (Born.). 
Fissurella mutabilis, Sowerby. 
Megatebennus scutellum (Gmelin). 



A. Regio Antarctica, including Gough Island. 

B. Ascension Island. 

C. Saldanha Bay, South Africa. 

A. — Regio Antarctica. 


Order Amphineura. 

Sub-order polyplacophora. 

Callochiton illuminatus (Reeve). 

Chiton illuminatus, Reeve, Conch. Icon., tab. xxii., fig. 147 (1847). 

C. (Callochiton) illuminatus, Smith, Proc. Zool. Soc. Lond., p. 35 (1881). 

Lepiclopleurus illuminatus, Rochebrune, Mission, Cape Horn, p. 141. 

Callochiton illuminatus, Tryon and Pilsbry, Man. Conch., vol. xiv., p. 51, pi. ix., figs. 92-94. 

Hab. — Trawl, Burdwood Bank, 56 fathoms, December 1, 1903 ; also Port William, 
Falkland Islands, Station 118, January 1903. 

A small, pale reddish or fawn-coloured species, the surface microscopically radiately 
striate and granulate, with the exception of the central arese, which are longitudin- 
ally sulco-striate. The Port William examples are slightly larger and of a darker 
fawn colour. 

Tonicia atrata (Sowb.). 

Cliiton atratus, Sowerby; Charlesworth's Mag. Nat.- Hist., 1840, p. 294; Conch. Illustr., 
figs. 57, 58. 

„ ,, Reeve, Conch. Icon., tab. xvii., fig. 103. 

Tonicia atrata (Sowb.), H. and A. Adams, Gen. Bee. Moll., i., p. 474. 

„ ,, „ Pilsbry, in Tryon's Man. Moll, xiv., p. 201, pi. xli., figs. 28-30. 

Hab. — Shore, Cape Pembroke, Falkland Islands ; and also at Port Stanley, 

A smooth, sometimes prettily banded or rayed species. Several very juvenile 
examples from contiguous localities are of similar pattern, but we dare not attempt 
their differentiation. 

Order Prosobranchiata. 

Sub-order diotocardia. 
Family Acmseidse. 

Acmsea ceciliana. Orbigny. 

Pat'l/a, cxilian'i, Orl>., Vuy. Amer. Merid., p. 482, tab. lxxxi., figs. 4-6. 

„ ,, Gay, Hist. Chili, viii., p. 260 (1854). 

.1 '■mxa ceciliana (Orb.), Tryon, Man. Conch., xiii., p. 33, pi. xxxiv., figs. 14-21. 


Hab. — Trawl, Port William, Falkland Islands, January 31, 1903. 

A small but attractive species. We have received it previously from the same 
locality. It was found, for instance, by Mr Rupert Vallentin,* and likewise by 
Miss Cobb, in Lively Island, in the enumeration of whose Mollusca t we at first named 
it textilis, Gould, a species confined to the western coasts of America. 

Family Patellidse. 

Patella senea, Martyn, var. deaurata, Gmel. 

Patella deaurata, Gmelin, Syst. Nat., xiii., p. 3703. 

Nacella senea, Mart., var. deaurata, Gmel.; Tryon, Man. Conch., xiii., p. 118, pi. xlvi., 

figs. 28-36. 
Patella xnea, Reeve, Conch. Icon. (Patella), fig. 9 (1855). 

Hob. — Stanley Harbour, Falkland Islands, shore. 

Patella fuegiensis, Reeve. 

Patella fuegiensis, Reeve, Conch. Icon., Patella, fig. 73 (1855). 

„ (Patinella) fuegiensis, Smith, Phil. Trans. Royal Soc. Lond., vol. clxviii., p. 180, 
tab. ix., figs. 14, 14a. 
Nacella fuegiensis, Tryon, Man. Conch., xiii., p. 121, pi. xlix., figs. 28-31 (1891). 

Hab. — On shore rocks, Hearnden Water, Falkland Islands, January 7, 1903. 
In somewhat puny condition, showing somewhat uneven ribs, and close latticed 

Patella polaris, Hombr. and J acq. 

Patella polar is, Hombron and Jacquinot, Ann. des Sci. Nat., II., xvi., p. 191 (1841). 

„ ,, v. Martens and Pfeffer, "Mollusken von Sud-Georgien," Jahrb. Hamburg. 

Wiss. Anstalt, iii., p. 101, tab. 2, figs. 11-13 (1886). 
Nacella polaris (H. & J.), Tryon, Man. Conch., xiii., p. 120, pi. xlix., figs. 21-27 (1891). 

Hab. — Dredged, Scotia Bay, South Orkneys, at 5-10 fathoms, January 2, 1904. 
Also at 1^ fathoms, December 3 and 18, 1903 ; and again at 6 fathoms, February 1, 1904. 

Station 325, Scotia Bay, South Orkneys. Dredged at 9-15 fathoms, December 17, 
1903, and various antecedent and subsequent dates. 

Section Zygobranchiata. 
Family Fissurellidse. 

Fissurella oriens, Sow. 

F. oriens, Sow., P.Z.S., 1834, p. 124; Conch. Illust, fig. 25; Thes. Conch., iii., p. 186, 
figs. 19, 20. Reeve, Conch. Icon., fig. 13. 

Hab. — Shore, Port William, Falkland Islands, January 31, 1904. 

* Journ. of Conch., vol. ix., p. 103 (1898). t lb., vol. x., p. 45 (1900). 



Fissurella picta (Gmel.). 

Un Lepas rare de Magellan, etc., Davila, Cat. Syst. et Raisonne des Cur., etc., i., p. 88, tab. iii., 

fig. 6, 1767. 
Lepas ovata ampla, etc., Martini, Syst. Conchyl. Cat., vol. i., p. 131, tab. xi., fig. 90. 
Patella picta, Gmel., in Syst. Nat., xiii., p. 3729. 
Fissurella picta, Gould, U.S. Expl. Exped., atlas, tab. 31, fig. 469, animal. 

Hab. — Dredged in 6 fathoms, Port Stanley, Falkland Islands, February 2, 1904. 

Tugalia antarctica, sp. n. (Plate, fig. 1). 

T. testa alba, oblonga, conico-depressa, parmophoroide, apice incurvo, postice subimpendente, antice^ 
obscure intus sulcato, superficie extus pulcherrime cancellata, striis tenuissimis, apud latera crassioribus, 
decorata, undique concentric^ delicate et arete striata, et liris incrementalibus irregulariter prsedita, intus alba, 
nitida, marginibus minute crenulatis. 

Alt. 7, long. 23, lat. 13 mm. spec. maj. 
„ 4, „ 16, „ 10 „ „ min. 

Hab. — Burdwood Bank, south of the Falkland Islands, at 56 fathoms. 

Allied to T. elegans, Gray, ossea, Gould, and parmophoroidea, Quoy, but of much finer 
sculpture and more convex. The internal groove leading to the sinus is, indeed, 
present, but very obscurely indicated, its region, however, being noted by the absence 
of longitudinal sculpture dorsally. 

Section Azygobranchiata. 

Family Trochidse. 

Photinula expansa (Sow.). 

Trochus expansus {Margarita), Sowerby, Conch. Illustr., figs. 16, 17. Philippi in Mart, and 

Chem., ed. ii., p. 254, tab. xxxvii., fig. 18. 
Trochus (Photinula) expansus, E. A. Smith, Phil. Trans. Roy. Soc, vol. clxviii., p. 167 

Photinula expansa (Sow.), H. and A. Adams, Gen. Rec. Moll., vol. i., pp. 427, 428 (1858). 

,, „ v. Jhering, in Nachrichtsbl. malak. Gesellsch. (1902), p. 99. 

Hab. — Station 118, trawl, Burdwood Bank, at 56 fathoms, December 1, 1903; also 
trawl, Port Stanley, Falkland Islands, at 1-10 fathoms, January 25, 1903. 

Photinula tseniata (Wood). 

Trochus txniatus, Wood, Ind. Suppl., pi. v., fig. 12. 
Margarita tseniata, Reeve, Conch. Icon., pi. xx., figs. 4, 4a. 

Photinula tseniata, H. and A. Adams, Gen. Rec. Moll., i., p. 427 (1858). v. Jhering in Nach- 
richtsbl. malak. Ges. (1902), p. 101. 

Hab. — Trawl, Port William, Falkland Islands, at 6 fathoms, January 1903. 


Photinula violacea (King). 

Margarita violacea, P. P. King, Zool. Journ., No. xix., London (1832), p. 346. 

,, ,, Hombr. and Jacquin., p. 59, pi. xiv,, figs. 29-31. 

Trochus violaceus, Philippi, Com-h. Cab., ii., p. 254, pi. xxxvii., fig. 19. 
Photinula halmyris, Rochebrune and Mabille, Miss. Gap Horn, Paris (1889): " Mollusques," 

p. 89, pi. iv., fig. 6. 
Photinula violacea, v. Jhering, in Nachrichtsbl. malak. Ges. (1902), p. 98. 

Hob. — Trawl, Burdwood Bank, at 56 fathoms. 

A very nacreous, pink-tinged species, with spire more or less elevated. 

Valvatella antarctica (E. Lamy). 

Margarita antarctica, Ed. Lamy, Exped. Antarctique Francaise (1907), p. 9, planche i., 

figs. 2, 3. 
, „ " Moll. Orcades du Sud," Bull. Mus. Hist. Nat, t. xii., p. 123 (1906). 

Hob. — Scotia Bay, South Orkneys, 9-10 fathoms, April 1903. Station 325, 
9 fathoms, August 1903. Macdougal Bay, South Orkneys, at 10 fathoms, December 
1903 ; also in shore pools in Scotia Bay. 

A smooth, Photinula-like species, but easily distinguished by the deep umbilicus. 
We follow Mr Edgar Smith (Nat. Antarct. Exped.: Nat. Hist., vol. ii., p. 11) in use 
of the generic name Valvatella, Gray (1857), and would refer to Proc. Mai. Soc, 
vol. iii. (1898), p. 205, and Journ. Conch., viii. (1897), p. 472, in justification thereof. 

Sub-order monotocardia. 

(a) Ptenoglossa. 

Family lanthinidse. 

Ianthina exigua, Lamarck. 

lanthina exigua, Lam., Anim. sans Vert., vi., p. 206. 

,, „ Tryon, Man. of Conch., p. 37, pi. x., figs. 17-22. 

Hob.— Station 72, tow-net, lat. 34° 2' S., long. 49° 7' W. 

Small but perfect and characteristic examples, pale, but finely striate. 

(b) Taenioglossa. 

Family Naticidae. 

Natica (Lunatia), sp. 

Ha b.— Surface-dredged by tow-net, lat. 9° 6' S., long. 31° 45' W. 

A few very minute, evidently juvenile forms. They are white, shining, globose, 
four-whorled, including the small, dark-coloured apical, hardly perforate, mouth large 
proportionately, operculum corneous, paucispiral. 


Family Capididse. 

Calyptrwa costellata, Phil. 

Calyptrxa costellata, Philippi, in Arch. Naturg. (1845), p. 62. 
Trochita corrugata, Reeve, Conch. Icon., xi., fig. 96 (1858). 

,, ,, Sowerby, Thes. Condi., v., p. 89. 

Calyptreea costellata, Phil.; H. Strebel, "Mollusken Fauna der Magalh. Prov., " Zool. Jahrb. 
Jena (1906), p. 159, Taf. xiii., figs. 88-97. 

Hab.— Trawl, 56 fathoms, Burdwood Bank, lat. 50° 25' S., long. 51° 0' W. 
All of the form clypeolum, Reeve, figured by Strebel, Taf. xiii., figs. 93 a, b. 

Crepidula dilatata, Lam. 

Crepidula dilatata, Lamarck, Anim. sans Vert., vii., p. 644. 

„ ,, ,, Sowerby, Thes. Conch., v., p. 65, figs. 100, 101. 

„ ,, „ Reeve, Conch. Icon., xi., 3. 

,, ,, ,, Hermann Strebel, "Mollusken Fauna der Magalh. Prov.," Zool, 

Jahrb. Jena (1906), p. 166, Taf. xiii., figs. 100, 101. 

Hab. — Trawl, Port William, Falkland Islands, 6 fathoms. 

An abundant species. The synonymy is very extensive. We have given it all, 
briefly, in Journ. Conch., ix., p. 101. 

Family Littorinidse. 

Littorina (Lasvilitorina) caligmosa (Gould). 

Littorina caliginosa, Gould, Proc. Boston Soc, iii., p. 83 (1849). 

Hydrobia caliginosa (Gld.), E. Smith, Phil. Tr. Roy. Soc. Lond., clxviii., p. 173, pi. ix., 

fig. 8 (1879). 
Lxvilitorina caliginosa (Gould), Pfeffer, in Mollusken von Siid-Qeorgien, p. 81, Taf. i., 

fig. 8, a-d (1886). 
,, ,, Pelseneer, Voy. "Belyica": Zool., Moll., p. 8. E. Lamy, "Moll, 

Orcades du Sud," Bull. Mus. Hist. Nat., t. xii. p. 112. 

Hab. — Trawl, Port William, January 31, 1904; and also Hearnden Water, 
Falkland Islands, January 7, 1903. 

Littorina (Lsevilitorina) coriacea, sp. n. (Plate, fig. 2). 

L. (Laevilitorina) testa pertenui, coriacea, periostraco corneo brunneo-olivaceo contecta, ovato-oblonga, 
haud nitente, anfractibus 5-6, quorum apicalis parvus, obtusus, cseteris apud suturas impressis, tumidulis, 
apertura fere rotunda, peristomate tenui, flexili, margine columellari paullum producto. 

Long. 7, lat. 3 mm. 

Hob. — Scotia Bay, South Orkneys. Dredged at 5-10 fathoms, January 2, 1904. 
More oblong than any form of L. caliginosa, Gould, and likewise larger than the 


several Georgia Island species described by Dr Pfeffer. # It is of remarkably thin 
substance, covered with a leathery brownish-olive epidermis, which extends over the 
outer lip, and to some extent also round the columellar margin, which is slightly 
basally produced. 

Littorina (Pellilitorina) pellita, v. Marts. 

Littorina pellita, E. von Martens, "Moll. Stid-Georgien," Sitz. Ges. Naturf. Fr. Berlin Jahrg., 

1885, p. 92. 
Laevilitorina pellita (Marts.), v. Martens and Pfeffer, "Moll. Stid-Georgien, '' Jahrb. Hamburg. 

Wiss. Anst., iii. Jahrg., p. 79, pi. i., figs, b, d (1886). 
Littorina {Pellilitorina) pellita (Marts.), E. A. Smith, "Southern Cross" Antarct. Exped., 

p. 204 (1902). 

Hob. — Shore pools, Scotia Bay, South Orkneys, December 6, 1903. Station 325. 
Off weed and stones at 9- 10 fathoms, May 1903. 

The largest example (alt. 15, lat. 12 mm.) comes from Scotia Bay. 

Littorina [Pellilitorina) setosa (Smith). 

Littorina setosa, E. A. Smith, Ann. and Mag. N. Hist., xvi., p. 69 (1875); Phil. Trans. Roy. 
Soc. Lond., vol. clxviii., p. 172, pi. ix., fig. 6. 

Hob. — Station 325, Scotia Bay, South Orkneys, 9-10 fathoms, June 1903. 

Much more uncommon than L. pellita, v. Marts. Our few examples are hardly 
typical, and in form show affinity with pellita. We are indebted to Mr E. A. Smith for 
having examined them. 

Lacuna divaricata, Fabr. 

Troc.hus divaricatus, Fabr., Faun. Grosnland., p. 392 (1780). 

Lacuna vincta, Forbes and Hanley, iii., p. 62, pi. lxxii., figs. 10-12; lxxiv., figs. 7, 8; 
lxxxvi., figs. 6-8. 

Hah. — Shore, Port William, Falkland Islands, January 31, 1904. 
We cannot dissociate the Antarctic specimens from those of Arctic or British 
shores. One example is banded, being probably the var. fasciata, Brown. 

Lacuna notorcadensis, sp. n. (Plate, figs. 3, 3a). 

L. testa parum rimata, ovato-rotunda, tenui, cinereo-alba, undiqne epidermide tenui straminea 
induta, anfractibus 4-5, quorum apicalis minutus, obtusus, lsevis, cseteris apud suturas gradatis, spiraliter 
acute penultimo quadri-, ultimo anfractu octo-carinato, lineis increnientalibus obscuris longitudinaliter prseditis, 
apertura fere rotunda, labro paullum effuso, coutinuo, columella fere recta. 
Long. 3, lat. 3 mm. 

Hob. — Station 325, Scotia Bay, South Orkneys, 9-10 fathoms, April 1903. 
Much resembling in general appearance, though not in substance and texture, a 
Fossarus, this little shell might possibly find a place near Lsevilitorina umbilicata, 
* Von Martens and G. Pfeffer, Mollusk. von Stid-Georgien, pp. 81 sqq. 


Pfeffer, which possesses slightly impressed carinse on the body-whorl ; or, with greater 
probability, in proximity to certain Lacunae, being nearly allied to L. wandelensis* 
Lamy, an Antarctic species recently described from Carthage Bay, Wandel Island. The 
substance is thin, covered with a pale and fugitive epidermis. The spiral keels on the 
body-whorl are eight in number, as against only five in L. wandelensis ; of these one 
is obscure, the others seem strong and well defined ; the spire likewise is more elevated. 

Family Rissoidee. 

Rissoa adarensis, Smith. 

Rissoa adarensis, E. A. Smith, Rep. "Southern Cross" Exped. Antarctic, Brit. Mus. (1902), 
p. 205, pi. xxiv., fig. 17. 
Nat. Ant. Exp., vol. ii., p. 8 (1907), pi. ii., fig. 2. 

Hah. — Station 325, Scotia Bay, South Orkneys, 9-15 fathoms. 

A few examples, almost typical. The author of the species considers the whorls 
rather less convex than in the type. 

Rissoa (Cingula) cingillus (Mont.). 

Turbo cingillus, Montagu, Test. Brit., p. 328, pi. xii., fig. 7. 
Rissoa cingillus, Forbes and Hanley, vol. iii., p. 122, pi. lxxix., figs. 9, 10. 
,, ,, Jeffreys, Brit. Conch., iv., p. 48. 

Hob. — Station 325, Scotia Bay, South Orkneys. Dredged at 9-15 fathoms, April 

We cannot separate the single example received from the European species, so 
similar is it in both form, size, and coloration. It may be that further supplies will 
show it to be a nearly allied southern species, though we are dubious as to this point. 

Rissoa edgariana, sp. n. (Plate, fig. 4). 

R. testa oblongo-fusiformi, vix rimata, parva, solidula, laevigata, anfractibus ad 7, quorum apicales duo 
niinuti, tumiduli, brunnei, ceteris apud suturas paullum impressis, subventricosis, pallide stramineis, flammis 
longitudinalibus fulvo-brunneis indistinctis ornatis, ultimo magnitudine caeteros exaequante, infra peripheriam 
versus basim spiraliter zonula fulvo-brunnea succincta, apertura ovata, peristomate tenui fer^ eontinuo, 
paullum expanso, columella paullum incrassata. 

Alt. 3-5, lat. 1-25 mm. 

Hah — Scotia Bay, South Orkneys, 9-15 fathoms. 

A brightly coloured, smooth, subhyaline species when in fresh condition. The 
majority, however, of those collected being defunct, are more solid in appearance. The 
markings are indistinct longitudinal brown flames surrounding the whorls ; and on the 

* Expffl. Antarct. Frang. Charcot, " Mollusques," par E. Lamy (1907), Bull. Mus. Hist. Nat. Paris (1905), 
p. 478, figs. 


body-whorl itself, below the periphery, there is a pale fulvous-brown spiral zone. We 
are indebted to Mr Edgar Smith for having examined this and other species, and 
have much pleasure in connecting his name with it. 

Rissoa fraudulenta, Smith. 
Rissoa fraudulenta, E. A. Smith, Nat. Ant. Exped., Nat. Hist., ii., p. 9, pi. ii., fig. 3 (1907). 

Hab. — Scotia Bay, South Orkneys. Dredged, 6 fathoms. 

Fine examples of this newly described, spirally striate Rissoa, differing from the 
types in being straw-coloured, otherwise identical. 

Rissoa parva (Da Costa). 

Turbo parvus, Da Costa, Brit. Conch., p. 104. 

Rissoa par va, Forbes and Hanley, iii., p. 98, pi. Ixxvi., figs. 2, 6; pi. lxxvii., figs. 6, 7 ; and 
pi. lxxxii., figs. 1-4. 
,, ,, Jeffreys, Brit. Conch., iv., p. 23 ; v., pi. lxvii., fig. 3. 

Hab. — Shore, Port William, Falkland Islands, January 31, 1904. 
These specimens cannot be dissociated from the typical European and British 
shell, and have probably been accidentally imported into the Falkland Islands. 

Rissoa (Onoba) Scotiana, sp. n. (Plate, fig. 5). 

R. (Onoba) testa albida, imperforata, solidiuscula, fusiformi, anfractibus 6, quorum apicales duo pervitrei, 
globulares, cseteris apud suturas impressis, ventricosis, undique arctissime spiraliter striatis, ultimo anfractu 
zonula straminea obscura ad medium prsedito, superficie interdum obscure longitudinaliter costellata, apertura 
semicirculari, intus albo-lactea, peristomate incrassato, continuo, paullum effuso, columella obliqua. 

Long. 3, lat. 1 mm. 

Hab. — Station 325, Scotia Bay, South Orkneys, 9-15 fathoms, April 1903. 

With a superficial resemblance to Rissoa (Onoba) striata, Mont., a well-known 
European species, this little shell seems, likewise, akin to R. (Ceratia) turqueti, 
E. Lamy, # recently described from Wandel Island, from which it differs in possessing 
one whorl more, in being not in the least degree rimate, in the thickened continuous 
peristome, and the obscure spiral straw-coloured zone encircling the body-whorl. 
Perhaps, in time to come, additional links to bind these two species together may 
be found. 

Rissoa, (Manzonia) zetlandica (Mont.). 

Turbo zetlandicus, Montagu, Trans. Linn. Soc, xi., p. 194, t. xiii., fig. 3. 
Rissoa zetlandica, Forbes and Hanley, iii., p. 78, pi. lxxx., figs. 1, 2. 
„ ., Jeffreys, Brit. Conch., iv., p. 20; v., pi. lxvii., fig. 1. 

Hab. — Shore, Port William, Falkland Islands, January 31, 1904. 

Obtained at the same time as R. parva (Da Costa), it is equally probable that this, 

* Bull. Mus. Hint. Nat. Paris (1905), p. 479, fig. ; Exped. Charcot, p. 6, pi. i., fig. 8. 


of which only one somewhat broken example occurred, is not a genuine native of these 
southern shores and seas. 

Eatoniella kerguelenensis, Smith. 

Eatonia kerguelenensis, E. A. Smith, Ann. and Mag. N. Hid., xvi. (1875), p. 70. 
Eatoniella kerguelenensis (Sm.), DalL, Bull. N.S. Nat. Mus., iii., p. 42 (1876). 

,, ,, E. Smith, "Zool. Kerguelen Moll.," Phil. Trans. Bog. Soc. Lond., 

clxviii., p. 174, pi. ix., fig. 40 (1879). 
Bissonia (Eatoniella) kerguelenensis, Smith, Man. Conch., ix., p. 391, pi. Ix., fig. 67 (1887). 

Hab. — Scotia Bay, South Orkneys, 9-10 fathoms, April 1903. Station 325, 
contiguous to first locality, 9-10 fathoms, July 1903 ; also shore pools in Scotia 
Bay, December 6, 1903. 

Quite typical, often encrusted with nullipore. Found likewise in South Georgia.* 
A further extension of range is now established. 

Family Litiopidse. 

Litiopa melano stoma, Bang. 

Buccinum litiopa, M. E. Gray, Figs. Moll. Anim., i., t. 24, figs. 1, 2. 

Litiopa melanostoma, Rang., Ann. des Sci. Nat., xvi., p. 303 (1829) ; Man. Hist. Moll., p. 129. 

„ bombix, Rang., I.e., p. 303 (1829). 

,, maculata, d'Orb., Moll. Cuba, ii., p. 149. 

Hab. — With tow-net, surface-dredging, and upon gulf-weed (sargassum) ; in several 
places, including particularly lat. 9° 6' S., long. 31° 45' W., December 16, 1902 ; and 
lat. 29° 54' N., long. 34° 10' \V., June 29, 1904. 

Very variable, and with a lengthened synonymy, of which only two or three have 
been given above. 

Family Cerithidse. 

Cerithium georgianum, Pfeft'er. 

Cerithium georgianum, Pfeffer ; von Martens and Pfeffer, Mollusken von Sud-Georgien, p. 97, 
Taf. ii., fig. 7(1886). 

Hab. — Scotia Bay, South Orkneys, 9-10 fathoms, April 1903. 

One example only, but in fine condition, entirely agreeing with Dr Pfeffer's 
plate and description of his South-Georgian species. It is a curious, small, strongly 
spirally ribbed and keeled shell, the carinse three in number on the lower whorls, four 
on the body-whorl itself. The apex is remarkably obtuse. 

* Pfeffer and von Martens, Vie Mollusken v. Siid-Georgien, 1886, p. 94, Taf. ii., fig. 5 a, b. 


Cerithium pvllum, Phil. 

Cerithium pullum, Philippi, in Arch. Naturg., 1845, p. 66. 

,, cedatum, Couthouy ; Wilkes, Expl. Exped., p. 148, fig. 174 a-d. Gould, Bost. Proc., 
iii., p. 123 (1849). 
Bittium cedatum, Couthouy, Mission de Cap Horn, p. 40. 
Cerithium pullum (Phil.), H. Strebel, " Beitr. der Mollusk. Fauna der Magalhaen Provinz," 

Zool. Jahrb. Jena, 1905, p. 652, Taf. xxiii., fig. 40 a-d. 

Hab. — Cape Pembroke, Falkland Islands, shore. 

Only dead and broken examples, but retaining the peculiar sculpture of this well- 
known species. 

Cerithiopsis mcdvinarum (Strebel, MS.), sp. n. (Plate, figs. 6, 6a). 

C. testa parva, imperforata, eleganter fusiformi, brunneo-castanea, anfractibus 7-8, apud suturas multum 
impressis, spiraliter fortiter bi-, ultimo tricarinulato, undique cancellatis et alveatis, apertura squarrosula, 
peristomate tenui, albescente, columella recta, paullum producta. 

Long. 4, lat. L5 mm. 

Hab. — Shore, Hearnden Water, Falkland Islands. 

Dr Hermann Strebel kindly examined our only example, and returned it with the 
above MS. name, requesting us to describe it. He informs us another example was 
found at Port Louis, by the Swedish South Polar Expedition, eight- whorled, not quite 
adult, in measurement 3 '9 x 1*4 x (aperture) 0*9 mm.; differentiated, but not yet 

Family Tritonidse. 

Gyrineum vexillum (Sowerby). 

Ranella vexillum, Sowerby, jr., Conch. Must., fig. 3. 

,, ,, Deshayes, in Lam. sans Vert., ix., p. 553, No. 30. 

,, „ Reeve, Conch. Icon., iii., t. 3, fig. 13. 

Triton ranelliformis, King, Zool. Journ., v., p. 347. 

Hah. — Gough Island, trawl, 100 fathoms. 

Only in juvenile condition. By some authors the Ranella vexillum, Sowb., is 
considered but a variety of the R. argus, Gmelin. 

(c) Gymnoglossa. 
No representatives of this section occurred. 

(d) Rachiglossa. 

Family Muricidse. 
Trophon brucei, Strebel. 

Trophon brucei, H. Strebel, "Beitr. der Mollusk. Fauna der Magalh. Provinz, : ' Zool. Jahrb. 
Jena, 1904, p. 230, Taf. viii., fig. 72. 



Hab. — Station 118, Port Stanley, Falkland Islands, trawl, at 1-10 fathoms, 
January 25, 1903. 

Near T. liratus, Couth., but of peculiar form. A larger series of specimens than we 
have seen is required to decide whether this be constant. 

Trophon cinguliferus, Pfeffer. 

Trophon cinguliferus, G. Pfeffer ; E. von Martens and G. Pfeffer, Mollusken von Siid- Georgien, 
Hamburg (1886), p. 70, Taf. i., fig. 2 a, b. 

Hab. — Dredged, 2-3 fathoms, Scotia Bay, South Orkneys, December 3, 1903. Also 
trap, Brown's Bay, South Orkneys, November 1903. 

Thus showing extension of geographical range. This species has a wonderful 
resemblance to certain forms of Purpura lapillus (L.). 

Trophon crispus (Couth.). 

Fusus crispus, Couthouy ; Gould, in Wilkes' Expl. Exped., p. 229, fig. 279 a-c. 
,, fimbriatus, Hupe ; Gay, Hist, de Chili, p. 165, pi. iv., fig. 7. 
,. ,, Smith, "Alert Surv.," P. Z. Soc. Lond., 1881, tab. iv., fig. 4. 

,, crispus, Herm. Strebel, "Beitr. der Mollusk. der Magalhaen Prov.," Zool. Jahrb. Jena, 
1904, p. 204, Taf. iii., fig 10 a-g. 

Hab.— Burdwood Bank, lat. 50° 25' S., long. 51° 0' W., 56 fathoms, trawl, 
December 1, 1903. Port William, Falkland Islands, trap, 6 fathoms, January 1903. 

Trophon geversianus (Pallas). 

Buccinum geversianum, Pallas; Spicil., Zool., fasc. 10, p. 33, pi. iii., fig. 1. Chemnitz, Conch. 

Cab., vol. iv., p. 130, pi. cxxxix., fig. 1297. 
Murex magellanicus, Gmelin, p. 3548, No. 80 (excl. var. /?). 

„ magellanicus and Murex peruvianus, Encyclop. Method.., vers., pi. ccccxix., figs. 4, 5. 

,, lamellosus, Dillwyn, vol. ii., fig. 730, No. 97. 

,, magellicanus, Wood, hid. Test., p. 132, pi. xxvi., fig 90. 

,, ,, Lamarck, Anim. s. Vert., vol. vii., p. 171, and (ed. Desh.) vol. ix., 

p. 589, sp. 46 (note). 
,, ,, D'Orbigny, Amerique meridionale, vol. v., p. 451, No. 392. 

,, patagonicus, D'Orbigny, Amerique meridionale, vol. v., p. 254, pi. lxii., figs. 2, 3, 

No. 393. 
,, varians, D'Orbigny, Amerique meridionale, vol. v., p. 452, pi. lxii., figs. 4-7, No. 394. 
Fusus geversianus, Reeve, Conch. Icon., vol. iv., pi. iv., fig. 2. 

Trophon geversianus, Sowerby, Thes. Conch.., pt 35, p. 59, sp. 1, pi. cccciv. (i. Gen.), figs. 7, 8. 
„ ,, Kobelt, in Conch. Cab. (ed. Kuster), pp, 275, 305, sp. 1, pi. lxxii., figs. 

1-3, and pi. lxxiii., fig. 1, pi. lxxv., fig. 1, pi. Ixxvi., figs. 1-2. 
,, ,, Tryon, vol. ii., p. 144, pi. xxxii., figs. 337-347, and pi. Ixx., figs. 

433, 435. 

Ilab.—Fort William, Falkland Islands, trawl, January 31, 1903. Scotia Bay, 
South Orkneys, trawl, 9-10 fathoms, December 17, 1903. 


Trophon hoylei, Streb. 

Trophon hoylei, H. Strebel, " Beitr. der Mollusk. Fauna der Magalh. Provinz," Zool. Jahrb. 
Jena, 1904, p. 227, Taf. viii., figs. 68 a-f, 69 a-c. 

Tlab. — Shore, Port William, Falkland Islands, January 31, 1904. 

We are not satisfied as to the specific distinctness of this species from T, liratus, 
Couth. The form and texture seem almost identical, but coloration, notably the 
dark purpurescent aperture, somewhat different. 

Trophon liratus (Couth.). 
Fusus liratus, Couthouy ; Gould, Wilkes' Expl. Exped., p. 231, fig. 282 a-c. 

Hob. — Station 118, Stanley Harbour, Falkland Islands. Dredged at 3^ fathoms, 
January 15, 1903. 

Many fine examples, in perfect condition, with operculum. We consider this the 
Buccinum cancellarioides, Reeve. It is evidently variable, and out of the forms 
Dr Hermann Strebel (to whom we are much indebted for the examination of our 
examples of this and its allies) has recently propounded several new species. 

Trophon minutus (Strebel, MS.), sp. n. (Plate, figs. 7, 7a). 

T. testa parva, fusiformi, albo-cinerea, anfractibus 5-6, quorum apicales 2 byalini, perlseves, cseteris 
apud suturas multum impressis, gradatulis, longitudinaliter arctissime costatis, costis lsevibus, incrassatis, 
numero ultimum apud anfractum circa 22, interstitiis lsevibus, apertura ovata, albo-lactea, labro paullum 
effuso, superne rectangulo, columella recta, canali paullum prolongato, lato. 
Long. 73, lat. 4, apertura 4 mm. 

Hab. — Station 325, Scotia Bay, South Orkneys, 9-15 fathoms, April-August 
1903, February 1, 1904. 

We are requested by Dr Hermann Strebel to draw up the description as above of 
a small Trophon, already differentiated by him in MS. as having been found in South 
Georgia Island recently by the Swedish South Polar Expedition. Only one example 
was dredged, not so perfect as our specimens, nor so large, having the apical whorls 
broken off, and but 3^- conserved whorls, measuring 37 x L9 mm., and numbering only 
17 costse on body- whorl, as against 22 in our type. On this account he suggests the 
varietal term major being applied to the South Orkney specimens, of which only two 
or three examples occurred, dredged from time to time in Scotia Bay. 

Trophon philippianus, Dunker. 

Trophon philippianus, Dkr. ; Mart, and Chem., ed. ii. (W. Kobelt), fol. 279, tab. lxxii., 
figs. 4, 5. 
,, geversianus (fusus), Gould, in Wilkes' Explor. Exped., p. 228. Var. philippianus, 

Dkr. ; H. Strebel, Zoologisch. Jahrb., p. 174 (1904). 


Hab. — Shore, and at 6 fathoms, Port Stanley, Falkland Islands. 
This agrees with specimens from Punta Arenas we have compared it with in the 
British Museum, and we are confirmed in our determination by Dr Strebel. 

Family Nassidse. 

Nassa (Ilyanassa) vallentini, sp. n. (Plate, figs. 8, 8«). 

N. testa ovato-oblonga, solidula, plicato-corrugata, plumbeo-brunnescente, anfractibus 7, quorum 2 
apicales plus minus laaves, caeteris irregulariter longitudinaliter crassicostatis, apud suturas magis incrassatis, 
impressis, aretissime" spiraliter liratis, liris rudibus, apertura rotundo-ovata, intus plumbeo-fusca vel brunnes- 
ceute, labro tenui, columella paullum excavata, baud callosa, canali ad basim breviter recurva. 

Long. 14, lat. 8 mm. 

Hab. — On the shore, Port William, Falkland Islands. 

To some extent resembling N. (Ilyanassa) ohsoleta, Say. — a common estuarine species 
on the Atlantic shores of the United States, — this Antarctic representative differs entirely 
in characters of sculpture, being conspicuous for its irregular thickened plicate costae, 
the whole surface being spirally sulco-lirate, mouth oval, leaden-coloured or red-brown, 
outer lip hardly thickened, columella slightly excavate, with no callus, canal basally 
shortly recurved. 

We name this Falkland Islands species after one who has in late years most 
thoroughly explored that group, our friend Mr Rupert Vallentin. 

Family Buccinidse. 

Chrysodomus (Sipho) archibenthalis, sp. n. (Plate, fig. 9). 

C. (Sipbo) testa fusiformi, tenui, albo-cinerea, undique epidermide pallide straminea tenuissima praedita, 
anfractibus, in specimine unico imperfecto, ad sex, quorum apicales duo leeves, submamillati, caeteris apud 
suturas impressis, tumidulis, spiraliter arete liratis, liris anfractus ultimi circa 12 suprA peripberiam, apertura 
ovata, columella paullum excavata, columella brevi. Operculo lutescente, unguiformi, corneo, nucleo ajucali. 

Long. 17, lat. 8 mm. 

Hab.—LoX. 62° 10' S., long. 41° 20' W., 1775 fathoms, trawl, March 10, 1903. 

Evidently not adult, and though this be the case, and the outer lip be to some 
extent broken away, we think this Sipho is worth describing, the upper whorls being 
very perfect. The ventricose, closely spirally lirate whorls seem characteristic. Only 
one specimen occurred, from the abysmal depth recorded above. It may be many years 
before another specimen is brought to light. We should imagine a full-grown shell 
would be at least twice the size — say 35 mm. in length. 

Chrysodomus (Sipho) crassicostatvs, sp. n. (Plate, figs. 10, 10a). 

C. (Sipho) testa parva, ovato-fusiformi, solidula, albo-cinerea, periostraco tenui olivaceo-stramineo con- 
tecta, anfractibus 5-6, quorum apicales 2-3 gradati, subhyalini, laeves, subtumidi, cgeteris spiraliter crassi- 
costatis, apud suturas impressis, ventricosis, ultimum apud anfractum numero 7, interstitiis laevibus, apertura 
ovata, labro'paullum effuso, columella excavata, ad basim breviter truncatula. 

Long. 6, at. 3 mm. 


Hab. — Station 325, Scotia Bay, South Orkneys, 9-10 fathoms, off weed and 
stones, April-August 1903. 

A very compact little species, of which a few examples, all precisely similar in size 
and sculpture, occurred. It is likely they are nearly, if not quite, adult, and are con- 
spicuous for the strong, acute spiral lirse or riblets encircling the three lower whorls in 
a uniform fashion, there being seven on the body-whorl. 

Neobuccinum eatoni (Sm.). 

Buccinopsis eatoni, Smith, Ann. and Mag. N. Hist., xvi., p. 68 (1875). 

Neobuccinum eatoni, Smith, Phil. Trans. Royal Soc, clxviii. (1879), p. 169, pi. ix., 

figs. 1, la. 
,, ,, ,, Try on, Man. Conrh., iii., p. 197, pi. lxxvii., figs. 357, 358. 

,, ,, ,, Watson, Rep. "Challenger" : Gastropoda, p. 216 (1886). 

E. A. Smith, "Southern Gross" Mollusca, p. 202 (1902). 
,. ,, ,, E. Lamy, Exped. Antarctique Frungaise (1906), p. 2. 

Hab. — Trawl, Scotia Bay, South Orkneys, 9-10 fathoms, December 17, 1903, and 
February 20, 1904. 

Several examples, mostly in live condition. The shell is of such thin texture, and 
the animal, perhaps, so large and swollen proportionately, that most examples were found 
to have their penultimate whorls broken, and the flesh protruding, when dredged up. 

Euthria fuscata (Brug.). 

Buccinum fuscatum, Bruguiere, Encycl. Meth., vers., p. 282 (1792). 

,, antarcticum, Reeve, Conch. Icon., iii., fig. 30 (1846). 
Tritonium schwartzianum, Crosse, Journ. de Conch., p. 171, tab. 6, figs. 9, 10 (1861). 
Euthria antarctica (Reeve), H. and A. Adams, Gen. Rec. Moll., t. i., p. 86 (1858). 

,, „ E. Lamy, "Gastr., Exp. Charcot," Bull. Mus. Hist. Nat., t. xi., p. 476 

,, fuscata (Brug.), Hermann Strebel, "Mollusk. d. Magalhaen Provinz," Zool. Jahrb., 

xxii., p. 611, pi. xxiv., figs. 69-72, 74-79 (1905). 
,, ,, E. Lamy, Exped. Antarct. Francaise : Gastr. et Pelecyp., p. 2 (1907). 

Hab. — Cape Pembroke, Falkland Islands, trawl, I— 10 fathoms, January 1903; also 
Port William, Falklands. 

Euthria magellanica (Philippi). 

Buccinum magellanicum, Phil., Abbildungen, vol. iii., p. 48, tab. i., fig. 14 (1848). 
Fusus rufus, Homb. and Jacq., Voyage de V " Astrolabe," vol. v., p. 107, tab. xxv., fig. 3 (1854). 
Euthria magellanica, Phil. ; H. Strebel, " Beitr. zur Kenntn. der Moll. Fauna der Magalhaen 

Prov.," Zool. Jahrb. (1905), p. 601. 

Hab. — Port William, Falkland Islands. 

Euthria michaelseni, Streb. 

Euthria michaelseni, H. Strebel, " Mollusken der Magalhaen Provinz," Zool. Jahrb. Jena, 
1905, p. 621, pi. xxi., figs. 6, 6a, Qb. 


Hab. — Port William, Falkland Islands, trawl, 6 fathoms, January 1903. 
An interesting new form ; conspicuous for a light spiral zone running round the 
centre of the body-whorl. 

Family Volutidae. 

Voluta (Cymbiola) ancilla (Sol.). 

Valuta ancilla, Solander, Portland Cat., p. 137, No. 1873. 
„ gracilis, Wood, Ind. Test., p. 209, Suppl., pi. iii., fig. 2. 
„ ancilla, Lamarck, Anim. s. Vert., vol. vii., p. 343, and (ed. Desh.) vol. x., p. 397, 

sp. 33. 
,, ,, Deshayes, Encycl. Method., vers., vol. iii., p. 1141, sp. 16, pi. ccclxxxv., fig. 3. 

,, ,, Kiener, Conch. Cat. (ed. Kuster), pp. 152, 153, pi. xxxii., fig. 4. 

,, ,, D'Orbigny, Amer. Merid., vol. v., p. 425, No. 333. 

,, mayellanica, Sowerby, Thes. Conch., vol. i., pt. 5, pi. liv., fig. 99. 
,, ancilla, Reeve, Conch. Icon., vol. vi., pi. xvii., fig. 39. 
,, magellanica, Gould, U.S. Expl. Exped., p. 278, pi. xx., fig. 357. 
,, ancilla, Crosse, "Cat. Voluta," Jour, de Conch., 1871, vol. xix., p. 299, sp. 61. 
„ (Cymbiola) ancilla, Kobelt, " Catalog Voluta," Jahrb. deutsch. maldk. Cesellsch., 1877, 

p. 310, sp. 61. 
,, ,, ,, Tryon, Manual, vol. iv., p. 97, pi. xxix., fig. 110. 

Hab. — Port William, Falkland Islands, trawl, 6 fathoms. 

Guivillea alabastrina, Wats. 

Wyvillea alabastrina, Watson, Prelim. Report, " Challenger" Exped., p. 12; Journ. Linn. Soc. 

Lond., vol. xvi., p. 332. 
Guivillea alabastrina, Watson, Rep. "Challenger" Exped.: Zool., "Gastropoda," vol. xv., 

pp. 262, 701, pi. xv., fig. 2. 
Voluta (Guivillea) alabastrina, Wats. ; Sowerby, Thes. Conch., iii., p. 304, pi. xviii., 

fig. 169. 

Hab.— Lat. 60° 10' S., long. 41° 20' W. Dredged at 1775 fathoms, March 18, 1903. 

One example in live condition, but hardly adult, and which is in shattered condition, 
with the larger portion of the body-whorl missing, the body protruding. The comparison, 
however, of the portion remaining with the figure in " Challenger" Report seems to 
indicate similarity as regards size. The whole substance of the shell is of papyraceous 

Section Toxoglossa. 

Family Couidse. 

Columbarium benthocallis* sp. n. (Plate, fig. 11a). 

C. testa eleganter fusiformi, tenui, papyracea, alba, epiderruide pallidc cinerea contecta, parum nitida, 
anfractibus ad 6-7, apicalibus . . . . 1 his ductus proximis erosis, quatuor ultimis fortissimo et ampliter 
apud medium carinatis, superficie undique tenuiter lira tula, et longitudinaliter irregularibus striis incremen- 
talibus praedita, ultimo anfractu subtus carina inferiori minus conspicua decorato, apertura subtriangulari, intus 
albo-cinerascente, peristomate tenui, margine columellari recto, canali longo, recto. 

Long. 45, lat. (carina anfr. ultimi inclusa) 18 mm. 

* B4v6os >c(£aoj, "beauty of the deep." 


Hab.— Trawl, lat. 62° 10' S., long. 41° 20' W., 1775 fathoms, March 10, 1903. 

A shell of papyraceous texture, as are so many abyssal species, whitish, with very 
thin ash-coloured epidermis. The apex of our only specimen is wanting, and the next two 
whorls much eroded ; the remainder are embellished with a central very acute and 
bold keel ; in the last whorl this carina is situate above the centre, and not far below the 
suture. A little lower comes a less-developed spiral keel. The triangular mouth is 
ashy-white within, the columellar margin straight, and the canal prolonged. 

Mangilia costata (Donovan). 

Murex costatus, Donovan, Brit. Shells, vol. iii., pi. xci. 
Pleurotoma costata, Don. ; Jeffreys, Brit. Conch., vol. iv., p. 379. 
Mangelia costata, Forbes and Hanley, iii., p. 485, pi. exiv., A, figs. 3-5. 

Hab. — Shore, Port William, Falkland Islands, January 31, 1904. 

One somewhat shattered example, but agreeing with authentic Mediterranean and 
British examples of the northern Mangilia costata (Don.), which is reported from South 

Family Cancellariidsp,. 

Admete magellanica, Strebel. 
Admete magellanica, H. Strebel, Zool. Jahrb., p. 594, Taf. xxii., figs. 29, 29 a-d (1905). 

Hab. — Port Stanley, Falkland Islands, shore, January 31, 1904. 
Allied to A. delicatula, Smith,* but with finer ribs. 

Order Opisthobranchiata. 

Sub-order tectibranchiata. 

Family Tornatinidse. 

Retusa truncatula (Bruguiere). 

Bulla truncatula, Brag., Encycl. Meth., p. 377 (1792). 

Utriculus truncatulus, Jeff., Brit. Conch., iv., p. 421, pi. xciv., fig. 2. Sars. Moll. Reg. Arch. 

Norv., p. 285, pi. xxvi., fig. 2; pi. xvii., fig. 18. 
Retusa truncate, Buq., Dantz., and Dollf., Moll. Rouss., i., p. 527, pi. lxiv., figs. 12-14. 
Bulla truncata, Ad. (non Gmelin), Tr. Linn. Soc, v., p. 1, pi. i., figs. 1, 2. 
Volvaria truncata, Brown, 111. Conch., G. B. ed., 1, pi. xix., figs. 17. 18. 
Cylichna truncata, Loven, Ind. Moll. Skand., p. 42. Forbes and Hanley, Hist. Brit. Moll., 

iii., p. 510, pi. cxiv., figs. 7, 8; pi. vv. fig. 4 (animal). Meyer and Mobius, Fauna der 

Kieler Bucht, i., p. 87 (animal). 
Bulla retusa, Maton and Rack., "Descr. Cat.," in Trans. Linn. Soc, viii., p. 128 (1804). 
Retusa obtusa, Brown, Pop. Encycl., ii., p. 78, pi. xvii., fig. 110. 
Volvaria pellucida, Brown, III. Conch., G. B. edit., i., p. 4, pi. xix., figs. 45, 46. 
Utriculus truncatulus, var. pyriformis, Monts., Nom. Gen. e Spec, p. 50. 

* Nat. Antard. Exped., Nat. Hist., vol. ii. (Brit. Mus. N.H.), 1907, p. 4, pi. i., figs. 5, 5a. 


Hab. — Shore, Port William, Falkland Islands, January 31, 1904. 
Indistinguishable from the European species. 

Section Siphonarioidea. 
Family Siphonariidse. 

Siphonaria redimiculum, Reeve. 

Siphonaria redimiculum , Reeve, Conch. Icon., ix., pi. v., fig. 21 (1856). 

„ ,, „ E. A. Hmith, "Moll, of Kerguelen," in Trans. Royal Soc. 

Lond., 1879, p. 16. 

Hab. — Hearnden Water, Falkland Islands, shore, January 7, 1903; also trawl, 
Port William, Falkland Islands, January 31, 1903 ; Gough Island, on the shore, 
April 22, 1904. 

Order Pulmonata. 

Sub-order basommatophora. 

Family Auriculidse. 

Marinula nigra, Philippi. 

Hab. — Gough Island, on the shore, April 22, 1904. 


Dentalium eujpatrides* sp. n. (Plate, fig. 12). 

D. testa magna, nitidissima, tend, alba, paullum arcuata, polita, lineis incrementalibus concentrice 
prsedita, versus apicem solum longitudinaliter obscure 1 multi-striata. Apice minimo, apertura rotunda. 
Long. 63, diametr. antic. 5, diametr. post. 1 mm. 

A fine, glistening white species, of thin and fragile substance. The concentric incre- 
mental lines of growth are frequent, and occasionally shown more plainly by staining 
of straw-colour. Towards the apex the surface is longitudinally closely striate, these 
striae extending for almost one-third of the total length. Some examples are straighter 
than others ; the majority are slightly gracefully arched. The aperture is circular, 
varying in diameter from 5 millimetres in the type, to 7 in one very broad specimen. 
From the figure and description t by Dr W. H. Dall, his D. callipeplum seems akin 
to our species. This was obtained in several stations in the West Indies during the 
Blake Expedition, at between 100 and 200 fathoms. 

* €U7roTp('87)s, patrician. 

t Bull. Mus. Cornp. Zool. Harvard, vol. xviii., p. 419, pi. xxvii., fig. 126 (1889). 


Dentalium shoplandi, Jouss. 
Dentalium shoplandi, Jousseaume, Bull Soc. Philomath, de Paris, vi., p. 102 (1894). 

Hob.— Lat. 71° 22' S., long. 16° 34' S., 1410 fathoms. 

Specimens are in the British Museum from " near Aden," dredged at 670 fathoms. 
We can see no difference in this southern form, either in size, texture, or sculpture. 
One example, dead, but characteristic. 

Order Protobranohiata. 

Family Nuculidse. 

Nucula minuscula, Pfeffer. 

Nueula minuscula, G. Pfeffer, in von Martens unci Pfeffer's Mollusken von Siid-Georgien, 
p. 128, Taf. iv., fig. 15 (1886). 

Hab. — Station 325, Scotia Bay, South Orkneys, 9-15 fathoms, April 1903. 

A single example only, for the determination of which we are indebted to the 
author, Dr Georg Pfeffer, of Hamburg. Its appearance in the South Orkneys is 
interesting, as considerably widening its geographical area. 

Yoldia eightsi (Couth.). 

Nucula eightsi, Couthouy; vide Jay, Cat. Shells, 1839, eel. iii., p. 113, pi. i., figs. 12, 13. 
Leda (Yoldia) eightsi, Hanley, in Sowerby's Thes. Conch., vol. iii., p. 142, pi. cexxx., fig. 164. 
Yoldia eightsi, Sowerby, Conch. Icon., vol. xviii., pi. v., fig. 26. 

Hab. — Station 325, Bay A, South Orkneys, at 9-10 fathoms, May 1903. Scotia 
Bay, at |- to 1^ fathoms, December 18, 1903 ; and again at 6 fathoms, February 1, 1904. 
Most examples were dredged on the latter occasion. 

Order Filibranchiata. 
Sub-order anomiacea. 

Family Anomiidse. 

Anomia ephippium, L. 

Anomia ephippium, Linn4, Syst. Nat., xii. ed. (1769). 

„ ,. Jeffreys, Brit. Conch., vol. ii., p. 30, pi. i., fig. 4. 

„ „ Smith, Report " Challenger" Exp. : Zool, xiii., p. 318. 

Hab. — Gough Island, trawl, at 100 fathoms. 

The specimens are all in young condition, and may possibly be an allied species ; 
but we can see no difference when compared with juvenile examples of this well-known 


ephippium. We are corroborated in our opinion by its having been noted in the 
Tristan d'Acunha group (Nightingale Island) during the Challenger Expedition.* 

Sub-order arcacea. 
Family Arcadse. 

Area (Bathyarca) strebeli, sp. n. (Plate, figs. 13, 13a). 

A. testa parva, orbiculata, paullum inaequilaterali, aequivalvi, alba, epidermide tenui, olivacea, fibrosa, 
praedita, radiatim tenuissime" arctilirata, lineis concentricis incrementalibus irregulariter cancellata, latere 
antico abbreviate, postico late rotundato, margine dorsali fere recto, deinde marginem apud ventralem leniter 
rotundato, umbonibus prominulis, obtusis, contiguis, conspicuis, dentibus parvis ad 18, pagini interna alba, 
marginibus lsevibus. 

Alt. 4 "5, lat. 5, diam. 4 mm. 

Hab.— Trawl, Station 291, lat. 67° 33' S., long. 36° 35' W., 2000 fathoms 
March 7, 1903. 

Allied to A. inaequisculpta, Sm. [Rep. "Challenger" Exp., xiii., p. 267, pi. xvii., 
figs. 8-8c), but differing from that species and its allies, A. pectunculoides, Scacchi, 
A. frielei, Jeffreys, and A. anaclima, Melv.,t in its rounder outline, more prominent 
umbones, and greater delicacy of texture. It is slightly larger than A. imitata, Sm. 
(I.e., p. 321, figs, in text), which seems a very variable form, dredged in the North Pacific 
Ocean, lat. 35° 22' N., long. 169° 53' E., at 2900 fathoms. This last is a coarser shell 
than A. strebeli, but its nearest congener, in our opinion. It gives us pleasure to 
connect with this interesting benthal species the name of Dr Hermann Strebel, who 
was good enough to examine it, and give us his opinion concerning it. 

Lissarca notorcadensis,\ sp. n. (Plate, figs. 14, 14a). 

L. testa parva, rotundo- vel ovato-trapezoide, solidiuscula, sordide" alba, asquivalvi, inaequilaterali. 
interdum epidermide fugaci, tenui, olivacea, pallida, partim tecta, undique concentrice arete striata, 
umbonibus obtusis, fere contiguis, ligamento partim externo, antice angusta, postice" expansa, marginem 
ad ventralem rotundata, pagina interna alba, laevigata, marginibus pulchre" crenulatis, dentibus ad 10, 
utrimque divergentibus. 

Alt. 5, lat. 6, diam. 2 - 75 mm. 

Hab. — Off weed, and attached to Bryozoa, etc., Station 325, Scotia Bay, South 
Orkneys, 9-15 fathoms; also in same bay, April-June 1903, June 1904. 

A proportionately thick, smoothish, white Lissarca ; some examples roundly, others 
(and more generally) ovately trapezoid ; larger as a rule than L. rubrofusca, Smith, 
which was found with it. The whole surface is closely concentrically striate. Within, 
the margins are crenulate ; the teeth, five on each side of the hinge-plate, divergent. 

Lissarca rubrofusca, Smith. 

Lissarca rubrofusca, E. A. Smith, Phil. Trans. Royal Soc. (1879), vol. clxviii., p. 185, pi. ix. 
fig. 17. 
,, ,, E. von Martens and G. Pfeffer, Mollusken von Sud-Georgien, 1886, p. 128, 

Taf. iv., fig. 14 a-e. 

* Rep. "Challenger" Exped., xiii., p. 318. t Proc. Zool. Soc. Lond. (1907), i., p. 794, pi. liv., fig. 6. 

\ v6Tos"OpicaS(s, from the locality. 


Hal.— Scotia Bay, South Orkneys, 9-10 fathoms, April 1903. Station 325, 
contiguous to first-mentioned locality, off weed, dredged August 1903. Likewise in 
shore pools, Scotia Bay, December 1903. All seemingly typical. 

Sub-order mytilacea. 

Family Mytilidse. 

Mytilus edulis, L. 

Mytilus edulis, Linne, Syst. Nat., xii. ed., p. 1157 (1769). 

„ „ Forbes and Hanley, ii., p. 170, pi. xlviii., figs. 1-4. 

„ ,, Jeffreys, British Conchotomy, ii., p. 104 (1863). 

„ „ Smith, Phil. Trans. Boy. Soc. Lond., vol. clxviii., p. 189 (1879). 

Hab. — Hearnden Water. Falkland Islands, 1903. 

This abundant European species is fast becoming almost universal outside the actual 
tropics. It was included by Mr E. A. Smith (vide the reference given above) as 
occurring in Kerguelen Island, and it is also reported from New Zealand. 

Mytilus magellanicus, Chemnitz. 

Mytilus magellanicus, Chem., Conch. Cat., vol. viii., pi. lxxxiii., fig. 742. 
,, ,, Reeve, Conch. Icon., vol. x., pi. vi., fig. 22. 

Hab. — Port William and Port Stanley, 6 fathoms, January 1904; also on shore, 
Cape Pembroke, Falkland Islands. 

Mytilus ovalis, Lamarck. 

Mytilus ovalis, Lam., Enc. Meth., pi. ccxix., fig. 3. 

„ „ Reeve, Conch. Icon., vol. x., pi. iv., sp. 14. 

Hab. — Hearnden Water, Falkland Islands, January 7, 1903. 

Philobrya meridionalis (Smith). 

Mytilus meridionalis, Smith, Beport, "Challenger" : Lamellibranchiata, vol. xiii., pp. 273, 
274, pi. xvi., figs. 3, 3a. 

Hab. — Station 325. Dredged at 9-15 fathoms, Scotia Bay, South Orkneys, 
May 1903. 

Originally described from a station between Kerguelen and Heard Islands, at 150 
fathoms, and, at about the same depth, from Prince Edward Island. 

Philobrya quadrata (Pfeffer). 

Philippiella quadrata, Pfeffer; E. von Martens and G. Pfeffer, Mollusk. von Sud-Georgien, 
1886, p. 119, Taf. iv., fig. 6 a-b. 

Hab. — Shore pools, Scotia Bay, South Orkneys. 


Philobrya wandelensis, E. Lamy. 
Phtlobri/a zoandelensis, Ed. Lamy, Exp. Ant. Franc. Charcot, 1903-5, p. 16, pi. i., figs. 15, 16. 

Hah. — Scotia Bay, South Orkneys. Dredged, 6 fathoms, February 1, 1904. 

Modiolarca mesembrina* sp. n. (Plate, figs. 15, 15a). 

M. testa parva, ovato-trapezoide, nigro-brunnea, tenui, convexa, aequivalvi, inaequilaterali, umbonibus 
prominulis, paullum incurvis, obliquis, interdum partim erosis, superficie omni epidermide nigresceiite contecta, 
sub lente radiatim longitudinaliter decorata, antice compressa, postice' rotundata, dente cardinali in utraque 
valva prsedita, pagina interna brannea. 

Alt. 4, lat. 4, diam. 2 mm. 

Hob. — Shore, Port Stanley, Falkland Islands. 

We cannot exactly match this with any of the known species. 

Order Pseudolamellibranchiata. 

Family Pectinidds. 

Pecten colbecki, Smith. 

Pecten colbecki, E. A. Smith, "Southern Cross" Ant. Exp. (1902), Brit. Mus. Publ., p. 212, 
pi. xxv., fig. 11. 

Hah— Lat. 64° 48' S., long. 44° 26' W., at 2485 fathoms, March 13, 1903. 

Fragmentary portions of the left valve only, the equal auricles being perfect, and a 
great part of the central superficies. This valve is flattened, and exhibits about fifteen 
ribs, somewhat uneven. There is uniform, close concentric stria tion. Colour, cinereous 
white suffused with dull rose-pink, with purplish tinge. Mr Edgar Smith described 
this species from a single right valve. The substance is so thin and papyraceous that it 
can be bent to a certain extent without injury. 

Pecten multicolor, sp. n. (Plate, figs. 21, 21a). 

V. testa delicata, mediocri, tenui, insequilaterali, sequivalvi, planato-convexiuscula, variis coloribus ornata, 
nunc flavida, nunc pallida violacea, interdum rufescente, vel alba, interdum pulchre" maculata, auriculis valde 
insequalibus, posticis radiantibus, radiis arete squamatis, anticis tenui-costatis, seque squamatis, valva sinistra 
costis ad 22, lsevibus, in statu juvenili pulclierrime squamatis, deinde laevissimis, paullum irregularibus, 
interstitiis 5-seriatim arctissime scobinatis, asperulatis, valva dextra costis arctis ad 50, interstitiis interdum 
planatis, interdum una serie squamarum scobinata decorata, umbonibus acutis, sublsevibus, pagina interna 
radiatim sulculosa. 

Alt. 21, lat. 19, diam. 5 mm. spec. maj. 
„ 16, „ 14, ,, 3 „ „ min. 

Hah. — Gough Island. 

A very delicate Pecten, variously coloured, in which the ribs and interstices of the 

* /juaeufSpii'os, southern. 


left valve, especially, exhibit much elaboration of sculpture. When young, round, shining, 
very fugitive scales exist, especially towards the ventral margin, whilst the interstices are 
closely scobinate, there being four or five closely imbricate rows of minute squamse. The 
right valve is by no means so elaborate, while the radiating ribs, irregularly placed, are 
more than double in number, not, indeed, leaving much space for interstitial development. 
Doubtless allied to P. limatula, Reeve,* and others of that section, the right valve, 
indeed, is hardly separable in sculpture from that of limatula ; but its fellow is more 
elaborately sculptured with close rows of interstitial scales than is the case in that 
species, which is reported from Nightingale Island, Tristan d'Acunha (Rep. " Challenger " 
Exped., xiii., p. 297, pi. xxi., figs. 5, 5<x). 

Pecten sp. 

A right valve of Pecten sp., the auricles very detrite, covered with small Balani, 
Membranipora, and other growths both without and within, but exhibiting about 35 
variable, on the whole distinct, clearly cut, narrow longitudinal ribs, may be a form of 
P. patagonicus, King. 

Hab. — Port William, Falkland Islands. Trawled at 6 fathoms. 

Pecten pteriola, sp. n. (Plate, figs. 16, 16a). 

P. testa parva, insequivalvi, insequilaterali, albo-cinerea, obliquante, valva dextra minore, haud nitente, 
concentric^ rudi-striata, aliter Isevi, auricula valvse hujus solum antica, lata, tenuisculpta, valva sinistra 
convexa, arctissime liris concentricis incrementalibus lamellatis prsedita ; interstitiis fere obtectis, umbone 
dextrse acuto, incurvo, margine dorsaliter utrimque sequali, recto, deinde apud marginem ventralem oblique" 
leniter rotundato. Pagiua intus alba, mtida. 

Alt. 4, lat. 4-5, diam. 1-5 mm. 

Hab. — Station 325. Dredged in Scotia Bay, South Orkneys, April 1903, at 9-1 0| 

The nearest approach to this particularly interesting little Pecten, of which several 
examples, all precisely similar, occurred, is P. aviculoides, Sm. (Rep. " Challenger" Exp. : 
Lamellibr., xiii., p. 325, pi. xxii., figs. 5, 5a). It is, however, not only double the size 
of that minute form, but also very differently sculptured, the ribs, so characteristic on 
the left valve of the Challenger species, which came from Prince Edward Island, being 
absent, and replaced by closely laminate concentric ridges or lirae. The specific name 
is chosen from the likeness to species of Pteria, Scop. ( = Avicula, Lam.) in miniature. 

Amussium octodecim-liratum, sp. n. (Plate, figs. 17, 17a). 

A. testa compressa, albo-lactea, subpellucida, tenui, rotunda, fere sequilaterali, inasquivalvi, valvis diverse 
sculptis, dextra pulcherrime cancellata, striis multis erectis concentric^ radiantibus, simul ac arctissime" longi- 
tudinaliter striatula, interstitiis quadratulis, valva sinistra delicate et arctissime concentrice striata, auriculis 
utrimque fere sequalibus, tenuisculptis, umbonibus ambobus acutis, subprominulis, pagiua interna alba, nitida, 
liris in valva utraque octodecim, ad margines extensis, prsedita. 

Alt. 9, lat. 11, diam. 2 - 5 mm. 

Hab.— Station 291, lat. 67° 33' S., long. 36° 35' W., trawl, 2500 fathoms. 

* Reeve, Conch. Icon., viii., pi. xxviii., fig. 124. 


A very thin, milky-white, delicately papyraceous Amussium, the valves almost 
equilateral, slightly unequal, auricles subequal, straight, finely sculptured ; the left valve 
beautifully concentrically radiately striate, the striae very close and fine ; and the right 
with finely cancellate upright striae, with quadrate interstices. Within, each valve 
exhibits an equal number of strong lirations, say 18, whence the specific name. 

Family Limidse. 

Lima (ManteUum) goughensis, sp. n. (Plate, figs. 18, 18a). 

L. testa alba, mediocri, convexa, obliquata, ovato-oblonga, sequivalvi, superficie omni longitudinaliter et 
regulariter multi-lirata, liris ad 48, interstitiis longitudinaliter sub lente per medium aequi- et unistriatis, 
lineis concentricis incrementalibus spiraliter praedita, umbonibus incurvis, margine dorsali fere recto, area 
ligamentari lata, centrali, pagina intus alba, longitudinaliter striatula, margine ventrali minute serrulato. 

Alt. 10, lat. 7, diam. 8 mm. sp. maj. 

Hah.— Trawl, 100 fathoms, Gough Island, April 23, 1904. 

A small species, if our specimens be adult. Although ordinary in form, with 
valves very convex and trapezoidly oblique, yet the delicate and regular longitudinal 
lirae, each with an accompanying single interstitial corresponding stria, seem 

Lima (Limatula) pygmsea, Philippi. 

Lima pygmeea, Philippi, Weigmann's Archivf. Naturgescli. (1845), p. 56. 

Limatula falklandica, A. Adams, Proe. Zool. Soc. Lond. (1863), p. 509. 

Radula {Limatula) pygmeea, Smith, Phil. Trans. Roy. Soc. (1879), vol. clxviii., p. 191. 

Lima (Limatula) pygmxa, Phil. ; Smith, Report " Challenger " Exp. ; Zool., xiii., p. 292. 

Hob.- — Station 325. Dredged in Scotia Bay, South Orkneys, at 9-10 fathoms, April 
1903 ; and again, South Orkneys, at 2-8 fathoms, December 1, 1903. Likewise in 
Macdougal Bay, South Orkneys. 

Order Eulamellibranchiata. 

Sub-order submytilacea. 

Family Lucinidse. 

Cryptodon falklandicus, Smith. 

Cryptodon falklandicus, A. E.Smith, Rep. " Oiallenger" Exp.: Zool., xiii., p. 190, pi. xiv., 
figs. 3, 3a (1885). 

Hub. — Scotia Bay, South Orkneys. Dredged at 9-10 fathoms, September 1903. 
In every way agreeing with the type. One specimen only, but perfect. This 
species has also occurred at Shallow Bay, Lively Island, Falklands * (Miss Cobb). 

* J. of Conch., ix., p. 105. 


Cyamium antarcticum, Phil. 

Cyamium antarcticum, Philippi, Arch. Naturg. (1845). 

„ „ H. and A. Adams, Gen. Bee. Moll, ii., p. 476, pi. cxiv., fig. 11 a, b 

Chemn., Man. de Conch., ii., p. 127, fig. 605. 

Hab. — Shore, Falkland Islands, January 1903. 
Only in young state, but perfect specimens. 

Cyamium falklandicum, M. & St. 
Cyamium falhlandicum, Melvill and Standen, Journ. of Conch., ix., pi. i., fig. 12 (1898). 

Hab. — Shore, Hearnden Water, Falkland Islands, January 1, 1903. 
Only in juvenile condition, occurring with the last-named species. The surface is 
mostly beautifully iridescent in these specimens. 

Family Erycinidse. 

Lassea consanguinea (Smith). 

Kellia consanguinea, E. A. Smith, Phil. Trans. Roy. Soc, vol. clxviii., p. 184, pi. ix., fig. 20 

Hab. — Hearnden Water, Falkland Islands, January 7, 1903 ; Scotia Bay, South 
Orkneys, 6-10 fathoms, February 1, 1904. 

Near L. miliaris, Phil., but apparently distinct. Varying somewhat in colour, 
from a uniform deep rose tint to straw-coloured, tinged with rose only at the dorsal 

Kellia cycladiformis (Desh.). 

Erycina cycladiformis, Deshayes, Trait, elem., pi. xi., figs. 6-9 ; P.Z.S. Lond., 1855, p. 181. 

Hab. — Burdwood Bank, trawl, 56 fathoms, December 1903. 

A species of wide distribution in southern latitudes. Keported from North Australia 
(Jukes), New Zealand (Quoy, Hutton, Hector), etc. 

Kellia lamyi. nom. nov. 

Kellyia australis, E. Lamy, " Moll, des Orcades du Sud," Bull. Mus. Hist. Nat. Paris, 1906, 
p. 124 (non Deshayes). 

Hab. — Scotia Bay, South Orkneys. Dredged at 6 fathoms. 

A minute species, which occurred gregariously in great numbers at the above station. 
Its obliquely ovate form, when seen with the valves closed, somewhat recalls a Nucula. 
In the left valve the single cardinal tooth is flanked by two very prominent incrassate 
laterals. In some specimens there is corrosion at the umbones. We have renamed 


this species in honour of M. Edouard Lamy, its original describer, the specific name 
he selected having been already employed by M. Deshayes. 

? Scacchia plenilunium, sp. n. (Plate, figs. 20, 20«). 

Sc. testa parva, elliptico-rotunda, convexa, nitida, albida, epidermide pallide olivacea, omnino contecta, 
laevigata, striis concentricis incrementalibus exceptis, apud unibones obtusos nequaquam prominulos, crassiore, 
marginibus fere rotundis, margiue dorsali intus pluteato, ligamento partim interno, dentibus lateralibus in 
valve utraque utrinque extensis, pagina interna pallide cinerea. 

Alt. 4, lat. 4'5, diam. 2 mm. 

Hab. — Shore, Cape Pembroke, Falkland Islands. 

A small, rounded, somewhat convex shell, with shining pale olive epidermis, smooth^ 
save for the concentric lines of growth. We are not sure of its genus ; but in some 
respects it possesses affinity to the European Scacchia elliptica, Phil. Within, the 
inner side of the dorsal margin protrudes as a kind of flattened narrow shelf or buttress 
in both valves, the extended lateral teeth being placed at either extremity of it. 
[Plenilunium, a full moon, from the rounded form.) 

Sub-order tellinacea. 

Family Tellinidw. 

Tellina (Msera) pusilla (Philippi). 

Tellina jms ilia, Pliil., Moll. Sic, i., p. 29, t. iii., fig. 9 a, b. 

„ pygmsea, "Phil." ; Loven, in Forbes and Hanley, i., p. 295, pi. xix., figs. 6, 7. 

Hab. — Shore, Gough Island, April 22, 1904. 

Only one half valve, but perfect, and coloured with rose radiations. We cannot 
separate it from the British species, which apparently has a wide range in the eastern 
tropics. It occurs, for instance, in the Persian Gulf, and on the Mekran Coast ; and 
this occurrence at Gough Island shows a further extension of its range southward in 
the Indian Ocean. 

Sub-order veneracea. 

Family Veneridse. 

Chione philomela (Smith). 

Venus philomela, E. A. Smith, Report " Challenger" Exp.: Lamillibr., Zool., xiii., p. 117, pi. ii., 
figs. 7, lb (1885). 

Hab. — Trawl, Gough Island, at 100 fathoms, April 23, 1904. 

Only small, dead, and imperfect examples. Originally described from Nightingale 
Island, one of the Tristan d'Acunha group, and within measurable distance of Gough 


Tapes [Amygdala) exalbida (Chem.). 

Venus exalbida, Chemnitz, Conch. Cat., vol. xi., p. 220, pi. ccii., fig. 1974. 
„ ,, Reeve, Conch. Icon., vol. xiv., fig. 13. 

,, ,, Sowerby, Thes. Conch., vol. ii., p. clxi., fig. 193. 

Chione exalbida, Deshayes, Cat. Conchif. Mus. Brit,, p. 154. 

„ (OmpJialoclathrum) exalbida (Chem.), Paetel, CataL, iii., p. 85. 
Tapes [Amygdala) exalbida (Chem.), E. A. Smith, Rep. " Challenger" Exped.: Lamellibr., xiii., 
p. 1 17. 

Hob. — Port William, Falkland Islands, trawl, January 31, 1904. 

Sub-order myacea. 

Family Glycimeridse. 
Saxicava arctica (L.), var. antarctica, Phil. 

My a arctica, Linne, Syst. Nat., p. 1113. 

Saxicava arctica (L.), Forbes and Hanley, i., p. 141, pi. vi., figs. 4-6. 
,, antarctica, Phil., Archivf. Nature/., 1845. 

Hob. — Grough Island, trawl, 100 fathoms, April 23, 1904; Burdwood Bank, 
trawl, 56 fathoms, December 1, 1903. 

We consider S. antarctica, Phil., by which name some would designate this protean 
species, merely as a synonym, or, at most, varietal, there being no salient characters 
which can be discerned by us as warranting their distinctness. 

Sub-order anatinacea. 

Family Lyonsiidze. 
[Lyonsia cuneata (Gray). 

Of this or a nearly allied species a single specimen, hardly adult, was found after a. 
gale in a stranded root of the giant Macrocystis, at Port Stanley, Falkland Islands, 
February 5, 1904. It has sometimes been considered an Entodesma or Mytilimeria.~\ 

Family Anatinidte. 

Ana Una elliptica, King and Brod. 

Anatina elliptica, King and Broderip, Zool. Journ., vol. v., p. 325. 
Reeve, Conch. Icon., xiv., fig. 14. 
Griffiths, Anim. Kingd., xii., pi. xxii., fig. 3. 

Smith, Rep. " Challenger" Exp.: LamelL, Zool., xiii. (1885), p. 76. 
„ "Southern Cross" Moll., p. 210, pi. xxv., figs. 9, 10. 
,, Nat. Antarct. Exped., vol. ii. (Brit. Mus. N.H.), p. 1, pi. iii., 
fig. 3 (1907). 
,, prismatica, Sowb.. P. Z. Soc. Lond., p. 87 (1834). 

Eab. — Station 325, Scotia Bay, South Orkneys. Dredged at 9-10 fathoms. 

A large, well-grown example. Smaller examples from the same locality, these last 



partially covered with bright green, somewhat shining epidermis. Mr Smith's figures of 
this species, references to which are given above, illustrate two extremes of form admirably. 

Order Septibranchiata. 

Family Uuspidariidas. 

Cuspidaria brucei, sp. n. (Plate, figs. 19, 19a and b). 

C. testa pro genere magna, ovata, teimi, papyracea, nitida, superficie omni delicate concentrice irregulariter 
lamellato-striata, insequilaterali, fere aequivalvi, convexa, umbonibus maxime contiguis, paullum inter se 
erosis, margine postico dorsaliter fere recto, deinde subrostrato, ventrali, cum margine antico, leniter rotundato. 
Pagina intus alba, lactea, nitida, dente cardinali in valva sinistra parvo, centrali, in dextra laterali lato, 
extenso. fossa ligamentari obliqua, parva. 

Alt. 17, lat. 26-5, diam. 10 mm. 

Hab.— Lat. 39° 48' S., long. 2° 33' E., 2645 fathoms. 

A very conspicuous, creamy-white, shining Cuspidaria, of papyraceous texture, of 
which only one example was dredged from the deep sounding given above. The 
posterior rostrum is abbreviate, the shell otherwise being gently rounded in the fore part. 
The umbones are close together, and detrite owing to friction. Within, the surface is pure, 
shining white, showing traces of the external concentric lamellar striation. We should 
place this species in the typical section of the genus {vide Smith, Rep. ''Challenger' 
Exited., vol. xiii., pp. 35-39 ; and Dall, Bull. Mus. Comp. Zool. Harvard, vol. xii., 
pp. 292 sqq.). To Mr W. S. Bruce, F.R.S.E., this beautiful addition to the genus is 

B. — At Ascension Island. 

Lotorium graudimaculatum (Reeve) [Tritori]. 

Hab. — Ascension Island, net, 18 fathoms, June 9, 1904. 
Now considered a var. of L. lotorium (L.). 

Nerita (Thelieosti/la) ascensionis, Gmel. 

Hab. — Ascension Island Harbour, dredge, 25 fathoms. 

Good examples of this usually worn species, in various stages of growth. 

C. — At Saldanha Bay, South Africa. 

Cli&topleura papilio (Spengler). 
Hab. — Two fine specimens from coaling jetty, Capetown Docks. 

Ischnochiton tigrinus (Krauss). 
Hab. — Shore, Hontjes Bay, Saldanha Bay. 


Chiton nigrovirens, Blainville. 
Hab. — Shore and trawl, Hontjes Bay and Saldanha Bay. 

Acanthochites Garnoti, Blainville. 

Hab. — Trawl, False Bay, May 8, 1904; Hontjes Bay, shore, and trawl; Saldanha 
Bay, shore and trawl. 

A fine and well-known species. 

Bullia annulata (Lam.). 
Hab. — Trawl, Saldanha Bay, 9-10 fathoms. 

Bullia laevigata, Chem. [Itevissima, Gmel.]. 
Hab. — Shore, Hontjes Bay, Saldanha Bay, May 20, 1904. 

Mytilus edulis, L., var. meridionalis, Krauss. 
Hab. — Trawl, Saldanha Bay, 25 fathoms. 

Mytilicardia (Thecalia) concamerata (Brug.). 
Hab. — Shore, Hontjes Bay, Saldanha Bay, 25 fathoms. 

Tapes pullastra (Montagu). 

Hab. — Trawl, Saldanha Bay, 25 fathoms. 

We cannot separate the several examples, all perfect, that were gathered at the above 
locality, from the European species. It is also recorded by Krauss, from Natal and 
Algoa Bay. # 

Oxy 'stele impervia, Menhe. 

Oxy stele tigrina, Chem. 
Hab. — Trawl, 25 fathoms, Saldanha Bay ; 0. tigrina also occurring at Hontjes Bay. 

Crepidula hepatica, Desh. 
Hab. — Trawl, 25 fathoms, Saldanha Bay. 

Gyrineum (Argobuccinum) argus (Gmel.). 

Hab. — Hontjes and Saldanha Bays, 5-15 fathoms. 
G. vexillum (Sowb.) seems synonymic. 

Phos plicosus, Dunker. 

This, which is the Nassa speciosa (A. Ad.), is a handsome species with certain 
characteristics of both genera, the Nassoid element, in our opinion, prevailing. We 
should therefore be inclined to restore it to the genus Nassa. 

* Vide G. B. Sowerby, Marine Shells of South Africa, p. 59 (1892). 


Hab. — Trawl, Reitz Bay, trap, 5 fathoms ; and in Saldanha Bay, 5 fathoms, 
May 20, 1904. 

Cominella limbosa (Lam.). 

Hab. — Trawl, Saldanha Bay, 25 fathoms; and shore, Hontjes Bay. 

Patella compressa, L., var. miniata, Born. 

We cannot dissociate this species from P. umbella, Gmelin. It is a species that 
varies its form in accordance with its habitat. 

Hab.— Trawl, 25 fathoms, Saldanha Bay, May 21, 1904. 

Patella (Scutellastra) granatina, L. 
Hab. — Shore, Hontjes Bay. 

Patella (Scutellastra) oculus, Born. 

With the last. These two handsome and similar species possess well - denned 
characters differentiating them from each other, but, though they usually occur 
together, have not yet been noticed to hybridise. 

Helcion pectinatum (Born.). 
Hab. — Shore, Hontjes Bay, Saldanha Bay. 

Fissurella mutabilis, Sowerby. 
Hab. — Trawl, 25 fathoms, Saldanha Bay. 

Megatebennus scutellum (Gmelin). 
Hab. — Trawl, Saldanha and Hontjes Bays. 


The following are the chief papers and treatises bearing on the Antarctic marine Molluscan fauna 
published from 1877 to 1907, inclusive: — 

1877. Smith, Edgar A., " Mollusca," in Zoology of the Transit of Venus Expedition, London, 1877, 4to, 

26 pp., pi. ix. [Kerguelen Island. Thirty-five species of marine Mollusca enumerated, and many 

1879. Smith, Edgar A., Kerguelen Island: the known Mollusca, 13 marine Bivalves, 26 Gastropoda, and 

1 Cephalopod enumerated, and their typographical and bathymetrical distribution pointed out : 

Phil. Trans. Roy. Soc. Lond., clxviii., pp. 157-192. 
1879. Studer, T., Arch.f. Nat., xlv., pp. 127-129. 
1885. Smith, Edgar A., "Report on the Lamellibranchiata collected by H.M.S. Challenger during the 

years 1873-76," in Reports on the Scientific Results of the Voyage of H.M.S. Challenger : Zoology, 

xiii., pt. xxxv., pp. 341, pi. xxv. 


1885. Smith, Edgar A., "Notes on the Lamellibranchiata," in Challenger Narrative, i., pp. 624-626, 

figs. 210-213. 

1886. Watson, R. Boog, "Report on the Scaphopoda and Gastropoda collected by H.M.S. Challenger 

during the years 1873-76," Zoology, xv., pt. xlii., pp. 756, pis. 1. [From the Messier Channel, 6 
species reported ; Straits of Magellan, 2 ; Falkland Isles, 13 ; Prince Edward Isles, 24 ; Kerguelen 
Island, 65; Heard Island, 16.] 

1886. Haddon, Alfred Cort, "Report on the Polyplacophora collected by H.M.S. Challenger during the 

years 1873-76," Zoology, xv., pt. xliii., pp. 50, pis. iii. [Lepidopleurus dalli, sp. n., Messier 
Channel ; L. dorsuosus, sp. n., Prince Edward Isles ; Leptochiton kerguelensis, sp. n., Kerguelen 

1887. Von Martens, E., and Pfeffer, G., "Die Mollusken von Siid-Georgien nach der Ausbeute der 

deutschen Station, 1882-83," Jahrbuch der hamburgischen tvissensch. Anstalt, iii., pp. 63-135, 
4 plates. [Many new species described by Dr Pfeffer.] 

1 887. Pfeffer, G., " Photinula ringei, nov. sp., von der Lemaire-Strasse," Verh. Ver. Hamburg, vi., 

pp. 98, 99. 

1888. Smith, Edgar A., " Diagnoses of New Species of Pleurotomidse in the British Museum," Ann. and Mag. 

N. Hist. (6), ii., pp. 300-317. [Description of PI. (Surcula) fuegiensis from the Straits of 

1889. Rochebrune, A. T. de, " Polyplacophora," Mission scientifique du Cap Horn, 1882-83: vi. Zoologie, 

pp. 131-143, pi. ix. [Twenty new Chitonidae described.] 
1889. Rochebrune, A. T. de, and Mabile, J., "Cephalopoda, Gastropoda, Lamellibranchiata," Mission 

scientifique du Cap Horn, t. c, pp. 1-129, pis. i.-viii. [Eighty species described, and several 

new genera.] 
1891. Smith, Edgar A., "Descriptions of New Species of Shells from the Challenger Expedition," P. Z. S. 

Lond., 1891, pp. 436-445, 2 plates. [Capulus eompressus, sp. n., South Patagonia; Turritella 

incolor, sp. n., Kerguelen Island.] 
1893. Hidalgo, J. G., Obras Malacologicas, pt. iii. : " Descripsidn de los Molluscos recogidos por la 

Comision cientifica eviada por el Gobierno Espanol k la America Meridional," Mem. Ac. 

Madrid, xvi., pp. 33-432. [Catalogues given of the Mollusca of the southern portion of South 

1895. Dautzenberg, Ph., " Liste de Mollusques du Chili," .4c;!. Soc. Chili, vi., pp. 64-67. 

1897. Chun, C, Die Beziehunyen zioischen dem arJctischen und antarldischen Plankton, Stuttgart, 8vo, 

64 pp., 1 map. [References to certain Antarctic pelagic Mollusca.] 

1898. Melvill, J. C, and Standen, R., "Notes on a Collection of Marine Shells from Lively Island, 

Falklands, with List of Species," J. Conch., ix., pp. 97-105, pi. i. (in parte). [Lachesis eutlirioides, 
Cyamium falklandicum, Thracia antarctica, spp. n.] 

1899. Thompson, DArcy W., "On a Supposed Resemblance between the Marine Faunas of the Arctic 

and Antarctic Regions," P. R. Soc. Edin., xxii., pp. 311-349. 
1899. Lonnberg, E., "On the Cephalopods collected by the Swedish Expedition to Tierra del Fuego, 1895- 

96," Scenska Exped. Mayellansldndern, ii., pp. 49-64, pis. iv., v. \_Gonatus antarcticus, Octopus 

patagonicus, spp. n.] 
1899. Stempell, W., "Die Muscheln der Sammlung Plate," Zool. Jahrb. Suppl, iv., pp. 217-250, pi. xii. 

[Aoicula (Meleagrina) magellanica, Teredo (Xylotrya) martensi, spp. n.] 

1899. Plate, L., "Die Anatomie und Phylogenie des Chitonen," Zool. Jahrb. Suppl., iv. (Fauna Chilensis), 

pp. 15-216. [Lepidopleurus medinse, Str. of Magellan, sp. n.] 

1900. Pilsbry, Henry A., "Species of Chlorostoma of Southern and Eastern Patagonia," Nautilus, xiii., 

pp. L 10-1 12. [Chlorostoma hidalgoi, C. orbignyanum, spp. n.] 

1901. Pelseneer, P., " Les Neomeniens de l'Expedition Antarctique beige, et la distribution geographique 

des Aplacophora," Bull. Ac. Belgique, 1901, pp. 279-292, et Congr. Zool., 1901, pp. 775 sqq. 
1901. Dall, W. Healey, "Synopsis of the Lucinacea and of the American Species," P. U.S. Mus., xxiii., 

pp. 779-833, pis. xxxix.-xlii. [Thyasira magellanica, sp. n.] 
1901. Preston, Hugh B., "Description of Cymbiola mangeri, a supposed New Species from the Falkland 

Islands," P. Mcdac. Soc. Lond., iv., p. 237, fig. 
1901. Melvill, J. C, and Standen, R.. "Mollusca collected by Mr Rupert Vallentin at Stanley Harbour, 

Falkland Isles, 1897-99," Journ. of Conch., x., pp. 43-47. 


1902. Smith, Edgar A., "On the Supposed Similarity between the Mollusca of the Arctic and the 

Antarctic Regions," P. Malac. Soc. Lond., v., pp. 162-166. 
1902. Smith, Edgar A., Report on the Collections of Natural History made in the Antarctic Regions during 

the Voyage of the "Southern Cross": Mollusca, pp. 201-213, pis. xxiv., xxv., London, 1902. 

[New species of Clione, Eatoniella, Lamellaria, Natica, Newnesia, gen. now, Paludestrina, 

Pecten, Philine, and Rissoa.] 
1902. Jhering, H. von, "Die Phot inula- Arten des Magellan-Strasse," Nachrichtsbf. deutsch. malak. Ges., 

1902, pp. 97-104. [Pliotinula lahillei, sp. n., Patagonia.] 

1902. Dall, W. Healey, "Illustrations and Descriptions of New, Unfigured, or Imperfectly Known Shells, 
chiefly American, in the U.S. National Museum," P. U.S. Mus., xxiv., pp. 499-566, pis. xxvii.-xl. 
[Toledonia perplexa, Str. of Magellan ; Antistreptus magellanicus, Str. of Magellan, gen. and 
spp. n. ; T rophon perelectus, Patagonia, sp. n.] 

1902. Dall, W. Healey, "Notes on the Giant Limas," Nautilus, xvi., pp. 15-17. [Lima patagonica, sp. n.] 

1902. Martens, E. von, "Einige neue Arten von Meer-Conchylien aus den Sammlungen der deutschen 

Tiefsee-Expedition," Sitzungsb. Ges. Natur. Berlin, 1902, pp. 237-244. [Puncturella analoga, 
sp. n., Kerguelen I.] 

1903. Bartsch, V., " A New Scissurella from Patagonia," Nautilus, xvii., pp. 90, 91. [Scissurella dalli, 

sp. n.] 

1903. Martens, E. von, "Die beschalten Gastropoden der deutschen Tiefsee-Expedition, 1898-99: 
A. Systematisch-geographischer Teil," Wiss. Ergebn. deutschen Tiefsec-Expcd. . . . Carl Chun, 
vol. vii., pp. 1-146, pis. i.-v. \_Lachesis australis, Odostomiopsis typica and circumrosa, spp. n., 
Kerguelen I. ; Scalaria instricta, Solariella periomphalia, spp. n., Neu Amsterdam.] 

1903. Pelseneer, P., Rcsultats du Voyage du S.Y. " Belgica" : Zoologie, " Mollusques (Amphineures, 
Gastropodes, et Lamellibranches)," Anvers, 1903, 85 pp., 9 pis. [Nov. sp. of Adacnarca (n. gen.), 
Bathyarca, Callocardia, Capulus, Cyamium, Cyclostrema, Dacrydium, Lxvilitorina, Leda, Lepto- 
chiton, Limojjsis, Margarita, Pecten, Philobrya, Rissoa, Scissurella, Sipho, Tergipes, Thards.~\ 

1903. Joubin, L., Resu/tats die Voyage du S.Y. " Belgica" : Zoologie, "Mollusques (Cephalopodes)," Anvers, 

1903, 4 pp. 

1904. Vallentin, R., "Notes on the Falkland Islands," Mem. Manchester Soc, xlviii., No. 23. 50 pp. 
1904-06. Strebel, Hkrmann, " Beitrage zur Kenntnis der Mollusken-Fauna der Magalhaen Provinz '' : — 

Part i. Zool. Jahrb. Syst. Jena, (1904), pp. 171-248, 6 plates. 
Part ii. „ „ (1905), pp. 576-666, 4 plates. 

Partiii. „ „ (1905), pp. 121-166, 1 plate. 

Part iv. „ „ (1906), pp. 91-174, 7 plates. 

[An exhaustive treatise on the Molluscan fauna of the Falkland Islands, with many new species described 
of Troplion, Pliotinula, Margarita, Calliostoma, Euthria, Savatipria, Thesbia, Beta, Drillia, 
Anoacme (gen. no v.), Admetc, Actseon, Utriculus, Columbefla, Bittium, and Euthria. Likewise 
Voluta martinsi, Natica anderssoni, Lamellaria fuegiensis, ampla, elata, magellauica, spp. nov. 
N.B. — This monograph is not yet complete.] 

1905. Eliot, Sir C. N. E., "The Nudibranchiata of the Scottish National Antarctic Expedition," Trans. 

Royal So<\ Edin., xli., pp. 519-532, figs. [Notxolidea gigas, purpurea, spp. n., South Orkneys; 
Tritonia appendiculata, sp. n., South Orkneys; Tritoniopsis brucei, sp. n., Gough Island.] 
1905. Smith, Edgar A., "On a Small Collection of Mollusca from Tierra del Fuogo," P. Malac. Soc. Lond.. 
vi., pp. 333-339, figs. [Chione fuegiensis, Mactra fuegiensis, Pliotinula. craiushayi, roseolineata, 
spp. n.] 

1905. Joubin, L., "Description de deux Elcdones provenant de ^Expedition du Dr Charcot dans 

PAntarctique," Mem. Soc. Zool. France, xviii., pp. 22-31, pi. iii. [E/edone charcoti, turqueti, 
spp. n.] 

1906. Laht, Iv, "Gastropodes Prosobranches recueillis par l'Expedition Antarctique Franchise du Dr 

Charcot," Bull. Mus. Hist. Nat. (1905), Paris, pp. 475-483, figs. [Cerithium charcoti, Lacuna 
wandelen&is, Rissoia turqueti, Turbonilla madrinensis, Margarita antarctica, spp. n.] 
1906. Lamy, E., " Lamellibranches recueillis par l'Expedition Antarctique Francaise du Dr Charcot," Bull. 
Mus. Hist. Nat. (1905), Paris, pp. 44-52. [Mactra laminifera, Montaguia charcoti, turqueti, 
Philobrya wandelensis, sublxris, spp. n.] 


1906. Lamy, E., Expedition Antardique Francaise (1903-1905) commandee par le Dr Jean Charcot: Mol- 
lusques, " Gastropodes et Pelecypodes," pp. 19, pi. i-, Paris (1906). [The species mentioned above 
redescribed and figured, with, in addition, Pecteji racovitzai, Pelseneer, sp. n.] 

1906. Vayssiere, A., Expedition Antardique Francaise (1903-1905) commandee par te Dr Jean Charcot: 
MoUusques : "Nudibranches et Marseniades," pp. 1-51, pi. iv. [Guyvalvoriafrancaisi, Charcotia 
granulosa, Rhodoplana loandeli, Archidoris tuberculata, Lamellariopsis turqueti, Marseniopsis 
antarctica, spp. n.] 

1906. Thiele, J., Expedition Antardique Francaise (1903-1905) commandee par le Dr Jean Charcot: 
MoUusques : " Amphineures," pp. 3. 

1 906. Thiele, J., Archeeomenia prisca, n. gen. n sp., mit 1 Tafel. 

1906. Thiele, J., Ueber die Chitonen der deutsche TieJ 'see- Expedition, mit 1 Tafel. 

1906. Thiele, J., Abdruck aus Wissensch. Ergebn. der deutsch. Tie/see- Exped. " Valdivia," 1898-99 (Carl 
Chun), Jena, 1906, pp. 317-336, Tafel xxix. 

1906. Lamy, E., "Sur quelques MoUusques des Orcades du Sud," Bulletin du Museum d'Histoire Nat. Paris 

(1906), pp. 121-126, figs. [Kellyia australis, Yoldia inxquisculpta, Vallettei, spp. n.] 

1907. Smith, Edgar A., National Antarctic Expedition, "Discovery," Nat. Hist., vol. ii., " Mollusca and 

Brachiopoda," pis. iii., pp. 1-12. [Neobuccinum tenerum, Troschelia, sp., Trophon longstaffi, 
coulmanensis, Thesbia? innocens, Admete delicatula, Amauropsis? rossiana, Neoconcha (n. gen.) 
vestita, Trichoconcha (n. gen.) mirabilis, Eulima conv^xa, Scala antarctica,, Rissoia fraud, ulenta, 
gelida, deserta, glacialis, Lovenella antarctica, Valvatella dulcis, crebrilirulata, refulgens, miuu- 
tissima, Lepeta antarctica, Bullinella gelida, Cheetopleura miranda, Cuspidaria tenella, Cardita 
antarctica, Kellia simulans, Tellimya antarctica, Cyamium denticulatum, Diplodonta incerta, 
Philobrya limoides, Limopsis grandis, hodgsoni, spp. n.] 


1. Tugalia antarctica. 

2. Littorina (Lsevilitorina) coriacea. 

3. 3a. Lacuna notorcadensis. 

4. Rissoa edgariana. 

5. ,, scotiana. 

6. 6a. Cerithiopsis malvinarum. 

7. la Trophon minutus. 

8. 8a. A 7 (Ilyanassa) vallentini. 

9. Clirysodomus (Sipho) archibenthalis. 
10, 10a. ,, „ crassicostatus. 
11a. Columbarium benthocallis. 

12, Dentalium eupatrides. 

13, 13a. Area (Bathyarca) Strebeli. 

14, 14a. Lissarca notorcadensis. 

15, 15a. Modiolarca mesembrina. 

16, 16a. Peden pteriola. 

17, 17a. Amussium ododecim-liratum. 

18, 18a. Lima (Mantellum) goughensis. 

19, 19a, 196. Cuspidaria brucei. 

20, 20a. 1 Scacchia plenilunium. 

21, 21a. Peden multicolor. 

tov. SocEdiiiF 



J.Green imp. 

( 159 ) 

VI.— The Pycnogonida of the Scottish National Antarctic Expedition. By Mr 
T. V. Hodgson, F.L.S. Communicated by Dr R. H. Traquair, F.RS. (With 
Three Plates.) 

(MS. received July 12, 1907. Read July 15, 1907. Issued separately January 24, 1908.) 

I am greatly indebted to Dr W. S. Bruce, the able leader of the Scottish National 
Antarctic Expedition, for the opportunity to describe the Scotia collections as regards 
the Pycnogonida and the Isopoda. We are now concerned with the Pycnogonida, 
and I regret that there has been so much delay before the production of the report. 
The collection is a large one, and extremely interesting — totally different from that made 
by the Discovery in the same region, but on the opposite side of the world. If smaller 
in the number of species brought home, in number of individuals it far exceeds that 
collection. Its principal interest lies in problems of distribution. 

First and foremost, there is the exceptionally interesting and important species 
Decolopoda australis, Eights (8), a species discovered and accurately described some 
seventy years ago, but forgotten, and, when first noticed, despised as a monstrosity or 
as a sample of defective work. For the present its relations must remain more or 
less conjectural. It is unquestionably a very close ally of the genus Colossendeis (12). 
Though it is early yet to make a positive assertion on the subject, it appears to share 
with that genus the capacity for depositing its eggs in some unknown hiding-place. 
The ova of both genera are unknown, and it is certain that in Colossendeis at least they 
are not carried by the male, or in fact by either sex. Decolopoda would appear to be 
abundant in the South Shetland and South Orkney Islands, and a second very closely 
allied species has been taken from the west coast of Graham's Land (Carthage Bay) by 
the French Antarctic Expedition. 

The second announcement of the discovery of a Pycnogonid with five pairs of legs, 
Pentanymphon antarcticum, was made by myself as one of the results of the Discovery 
Expedition (10). In the present collection there is only a single individual of this 
species, and that an adult female ; but it has also been captured by the French and 
German Expeditions. It may be said to have a circumpolar distribution, and lives at 
a depth of from 10 fathoms to just over 200. 

For the rest of the collection, there are several new species, none of which, however, 
can be placed in new genera. 

Palleno-psis has a new species very closely allied to the P. pilosa of Hoek, but 
distinguished by spurs on the lateral processes and legs. Dr Hoek's species is, more- 
over, a deep-water form, while this one comes from inside the 10-fathom line. 

Nymphon contains three species from the Antarctic, and another, a new species, from 


the African coast. Of the Antarctic species, two are identical with species taken by 
H.M.S. Challenger off New Zealand in 1100 fathoms (Station 168). N. longicoxa was 
taken in 1410 fathoms, lat. 71° 22' S., long. 16° 34' W. ; and N. co?npactum was taken 
in 1775 fathoms, lat. 62° 10' 5" S., long. 41° 20' W. The great distance from the point 
of their original discovery is well worthy of note, as well as the increase of depth. 

Chwtonymplwn contains four species with some interesting facts in distribution. 
C. orcadense is described as a new species, and was taken in Scotia Bay in immense 
numbers, and a single specimen from the Burdwood Bank, off the Falkland Islands. 
At first sight it is like C. australe, Hodgson, but is readily distinguished by the distinct 
auxiliary claws. C. assimile is a new species of which there are many specimens in the 
collection. This one and C. australe and C. altioculatum, Mobius, are all very closely 
allied. In my report on the Discovery collection I stated that I was unable to regard 
C. altioculatum as distinct from C. australe. C austrinorum I regarded as a variety 
of C. australe, though a well-marked one. These four species or varieties, whichever 
they may be, differ more definitely in their setose covering than in any other particular. 
I have no reason to modify my opinion on their specific differences ; but with regard 
to C. altioculatum, Mobius, there is most room for doubt. Examination of the type 
specimens reveals the fact that there is a fringe of small setae at the posterior border of 
the trunk segments. This feature is not very prominent, and is not brought out by 
Professor Mobius' description ; and if this is not to be regarded as a specific character, it 
has a certain value when the variation and distribution of species are under consideration. 
All the other peculiarities of the animal agree closely with those ascribed to C. australe, 
and one does not feel justified in separating the two as distinct species. No specimens 
of C. australe or this particular variety occur in the Scotia collection, but a very 
close relation exists in Chsetonymplion assimile, which is described as a distinct species. 
It is readily distinguished by the setose character of its body, which is devoid of setae, 
except for a few long ones as a distal fringe on the lateral processes ; other features 
show that it may be regarded as distinct, though closely allied. 

An important bond of union between all these species is the tall ocular tubercle, 
which caused Professor Mobius to derive the name of his species from that organ. One 
specimen of C. assimile and another of C. orcadense have two fairly well-developed 
eyes on one side of the tubercle, in addition to the four at the summit. 

A purely sexual feature lies in the fact that in C. australe, C. austrinorum, C. alti- 
oculatum, and C. assimile the ovigers of the adult male all have the fifth joint abruptly 
expanded to more than double its diameter for its distal half; the following joint also 
is considerably enlarged, but most so proximally. Another is the enormously developed 
distal fringe which occurs ventrally on the third coxae of the two posterior pairs of legs 
in the male. 

C. mendosum is not uncommon in Scotia Bay at 10 fathoms; at the other side of 
the hemisphere, and farther south by 1 6°, the Discovery found it equally abundant, but 
only below the 100-fathom line. 


C. brevicaudatum, Miers, only previously known from Kerguelen, now extends its 
range to Scotia Bay. 

Ammothea. — Two individuals of a single species of this genus were found, and are 
identified with Professor Bouvier's^I. communis, which was found in great numbers by 
the Francais on the west side of Graham's Land. The members of this genus appear 
to be abundant in the Magellan Province, but none were seen by the Discovery in 
M'Murdo Sound. 

Leionymphon is a genus instituted by Professor Mobius for some immature 
specimens found off Bouvet Island. The collection of the Discovery, which was rich in 
these forms, necessitated a revision of the genus, which now includes no less than eight 
species, a key to which is given in my report on that collection (13). The two species 
included in the present collection were described many years ago by Dr G. Pfeffer, 
under the generic name of Ammothea. They were then recorded from South Georgia, 
and it is only to be expected that they should occur also in the South Orkneys. L. 
grande, however, has had its range extended considerably, and evidently has a circumpolar 

Colossendeis is represented by two individuals, two very distinct species claiming 
them. One is a Challenger species, C. leptorhynchus, and was taken in the area where 
it was first found. The other is introduced as new ; it is quite blind — not an unusual 
character of the genus, but the more surprising as it is a shallow-water species. 

When considering geographical problems during our stay in the Antarctic regions, 
I accepted the mean annual isotherm of 45° F. for the ocean surface, as defined by 
Dr A. Buchan (4) in the concluding volume of the " Challenger " Reports, as the northern 
limit of the Sub- Antarctic region, partly perhaps as a matter of convenience, and partly 
because it is a natural limit which includes Kerguelen Island and its neighbours, which 
have, for a long time, been regarded as " Antarctic," and which have become " classical ' 
ground by the work of that expedition. The value of this boundary is emphasised by 
the fact that Professor Pelseneer (23), when reporting on the Mollusca of the Belgica 
Expedition, and examining the subject exhaustively, fixed on the isotherm of 40° F. for 
the air in July, and a similar isotherm for the ocean surface, but a minimum and not a 
mean temperature ; for this latter isotherm Sir John Murray is the authority. All 
the three isotherms above quoted are in very close accord, and a little south of lat. 45° S. 
on the Pacific side, a little above it on the side of the Atlantic and Indian Oceans. As 
to the boundary between the Sub- Antarctic and the Antarctic regions, I suggested in 
my report on the Pycnogonida of the Discovery that lat. 60° S. might be provision- 
ally regarded as such ; it includes all the glaciated lands of the Antarctic continent and 
the islands connected therewith. A more satisfactory limit would be the average limit 
of pack ice, if a surface phenomenon is to be accepted ; otherwise the centre of the 
trough between the Antarctic continent and the more northern lands would make a 
natural division, but the position of this trough will for some time at least be a matter 
of conjecture. 



Iii a recent paper on the Pycnogonids of the Magellan Straits (14) it became 
necessary to define that area, and I suggested the division of the Antarctic and Sub- 
Antarctic regions into three provinces corresponding with the continents from which the 
attack on those regions can be made. The limits are as yet purely arbitrary, being 
defined by lines of longitude which may not hold good when our zoological knowledge of 
these regions becomes greater than at present. The division suggested was as follows : — 

(1) Magellan Province, long. 20° W. to long. 130° W. 

(2) Australasian Province, long. 130° W. to long. 100° E. 

(3) Kerguelen Province (African), long. 100° E. to long. 20° W. 

The following list contains all the known species of Pycnogonida from the Antarctic 
and Sub- Antarctic regions — a list which contains some seventy-two species, thirty-four 
of which belong, as far as at present known, exclusively to the Antarctic area. Thirty 
of them belong to the Sub-Antarctic region, and of these, seven find their way farther 
northward. Those marked with an asterisk are contained in the Scotia collection. 


Pycnogonum magellanicum, Hoek 
„ magnirostre, Mobius 
Phoxichilus australis, Hodgson . 
Pallene dimorp/ta, Hoek 
Pseudopallene cnrnigera, Mobius 
„ australis, Hodgson 

Pallenopsis patagonica, Hoek 
„ pilosa, Hoek 
„ Jiuminensis, Kroyer 
,, glabra, Mobius 

,, villosa, Hodgson 
,, hiemalis, Hodgson 

„ lanata 

Anoplodactylus neglectus, Hoek 

,, petiolatus, Kroyer 

Nymphon gracile, Leach 
,, gracilipes, Miers 
„ brachyrhynchum, Hoek 
„ hamatum, Hoek . 
,, fuscum, Hoek 
,, meridionale, Hoek 
,, antarcticum, Pfeffer 
,, hiemale, Hodgson 
,, lanare, Hodgson. 
,, adareanum, Hodgson 
!> frigidum, Hodgson 
,, tridentatum, Hodgson 
,, longicoxa, Hoek . 
„ compaction, Hoek 
,, articular e . 
' Clixtonymphcn brevicaudatum, Miers 
villosum, Hodgson 
biarticulatum, Hodgson 
rnendosum, Hodgson 
austrah, Hodgson 
,, var. austrin 



* Chxtonymphon orcadense . 

* „ assirnile 

* Pentanymphon antarcticum, Hodgson 
Leionymphon striatum, Mobius . 

grande, Pfeffer 
gibbosum, Mobius . 
minum, Hodgson . 
Clausi, Pfeffer 
australe, Hodgson . 
glaciate, Hodgson . 
spinosum, Hodgson 
Ammothea Hoeki, Pfeffer . 

„ Wilsoni, Schimkewitsch 

* ,, communis, Bouvier . 
,, curculio, Bouvier 

Tanystylum styligerum, Miers 
,, Dohrnii, Pfeffer 

„ Chierchix, Schimkewitsch 

,, longicaudatum, Hodgson 

Austrodecus glaciale, Hodgson . 

Austroraptus polar is, Hodgson . 

Ascorhynchus glaber, Hoek 

Rhyncothorax australis, Hodgson 

Colossendeis gigas, Hoek 

* ,, leptorhynchus, Hoek 

„ gigas leptorhynchus, Hoek 

,, megalonyx, Hoek 

,, robust a, Hoek . 

„ gracilis, Hoek . 

,, australis, Hodgson . 

,, glacialis, Hodgson . 

,, frigida, Hodgson 

„ rugosa, Hodgson 

„ mage/lanica, Hodgson 

* ,, orcadensis 

* Decolopoda australis, Eights 

„ antarctica, Bouvier . 




Other species taken on the voyage, but from the African coast only : — 

Nymphon capense. Discoarachne brevipes, Hoek. 

Palleiiopsis lanata (Plate II., figs. 4, 4a). 

Specific Characters. — Body well built, with lateral processes widely separated, each with 
a prominent spur distally and dorsally ; body and appendages richly setose. 

Ocular tubercle conical, placed in front of cephalon ; with anterior eyes large, posterior 
ones much smaller and above them. 

Oviger ten-jointed, with simple setae only ; no terminal claw, differing in the two sexes. 

Legs with long terminal claw and small auxiliaries. Spurs on the femora and first 
tibiae distally and dorsally. 

The Body is fairly robust, with widely separated lateral processes ; these are not very 
long, slightly dilating, with a strongly developed spur distally, as well as a fringe of 
long setae. The segmentation is distinct ; each segment carries a fringe of long setae 
along its posterior border. 

The Cephalon is long, cylindrical, and carries the ocular tubercle at its anterior 
extremity. This rather peculiar structure, seen from behind, is stout, conical, and of 
some considerable elevation ; seen laterally, it is still conical, but the front upper part of 
the cone is incurved above the anterior pair of eyes, which are very large ; the posterior 
pair are scarcely half the size, behind and above them. 

The Abdomen is long, almost cylindrical, being slightly dilated just beyond the 
middle. It is directed upwards to some extent, not articulated to the trunk, and pro- 
vided with a number of fine long setae. 

The length of the body is 10 mm., the trunk only 7 mm., and its extreme width 5 mm. 

The Proboscis is nearly as long as the scape of the chelifori, and rises ventrally 
behind the position of the ocular tubercle. It is cylindrical and completely covered 
with stiff setae, which are largest and most conspicuous ventrally and distally. 

The Chelifori are well developed, and rise from the anterior border of the cephalon 
just underneath the ocular tubercle. The scape is long and two-jointed, the joints being 
subequal in size and covered with rather short stiff setae, which are more conspicuous 
as distal fringes. The chela has small fingers turned inwards almost at a right angle 
to the palm ; the movable one is much the larger, and is without a setose pad at its base. 
Neither bear teeth. The palm itself is not quite so long as a joint of the scape, and is 
covered with shorter and stiff setae. 

The Palps are reduced to stout, rounded, setose buttons ; they rise laterally close to 
the proboscis. 

The Ovigers are ten-jointed, and differ very considerably in the two sexes. In the 
male, the first joint is very small, the second is very much longer and considerably 
dilated distally ; the third is not half as long as the second, and is articulated at an 
angle to it, and not in direct continuation ; its outer border is rounded, the inner one 


being straight ; these two joints bear setae of varying length on their outer margins. 
The fourth and fifth are long and slender, slightly curved, the fifth being a little the longer 
and dilated distally ; both are rather scantily covered with short stiff setae, and the 
fourth has a large glandular (?) opening near its base. 

The fifth joint is quite short and very stout, bent rather than curved, with a thick 
covering of long setae on its outer margin. Of the remaining four, which curve in the 
opposite direction to the preceding joints, the first two are subequal in length, the two 
last taper, the terminal one being a little the shortest ; all these are richly clothed with 
long setae of quite a simple character. 

In the female this appendage is shorter and quite different in structure. The first 
joint is very short ; the second is about twice as long and dilated distally ; the third is 
about as long as the first, and articulated to the second in a normal manner. The fourth 
and fifth are longer, the fourth being much stouter and a little longer than the fifth, and 
about two-thirds the length of similar joints in the male. All these joints are setose, the 
setae being small at first, increasing in size and number to the extremity of the fifth. 
The sixth is scarcely half the length of the fifth ; the seventh, again, much smaller ; the 
eighth is longer, and the terminals taper, but differ little in length. All these are 
plentifully clothed with rather long simple setae, and chiefly on the ventral side. 

The Leg extends to about 40 mm. Of the three coxae, the first is the shortest, the 
second as long as the other two together ; the proportions of the three following joints 
are as 9, 8 "5, 10. The tarsus is very small, the propodus slightly curved, with a slender 
terminal claw nearly as long, and two small auxiliaries. The entire appendage is thickly 
covered with setae ; these, on the dorsal surface more especially, are long and slender ; 
those on the ventral surface are much shorter. In addition to these there are short and 
fine setae distributed uniformly over the whole limb. As usual, the setae on the second 
tibiae become more spinous distally, and the distal fringe of this joint is distinctly 
spinous ventrally. The ventral surface of the tarsus bears spinous setae which become 
very prominent distally. The propodus bears three stout spines proximally and 
ventrally, the remainder of the surface being occupied by a band of spinous setae. 
There is no projecting heel, but a well-developed distal fringe projects over the base 
of the claw. The femora and first tibiae of both sexes bear a prominent spur distally 
and dorsally. The lateral line is distinct. 

The Genital apertures of the male occur on conspicuous swellings of the second coxae 
of the two posterior pairs of legs. About the middle of the femur, ventrally, there is 
a slight enlargement extending for some distance along the joint ; from the middle of 
this projects the tubular process so characteristic of the males of this genus. In this 
case its length is 7 mm. 

The female is larger, and the conspicuous genital apertures open on the second coxae 
of all the legs, and the distal extremities of these joints are much dilated in consequence. 

The females appear to be larger than the males. 

Eleven specimens of this species were taken in Scotia Bay at a depth of L4 fathoms, 


March 1903. One of the males is carrying a large number of young. Some of these, 
recently hatched, have a very stout, massive body bearing no resemblance to the structure 
of the adult. The proboscis and chelifori are very distinct ; two pairs of small append- 
ages lie behind these. They show a segmentation into joints. Half way between these 
and the posterior extremity of the body-mass is another pair of appendages, stout 
stumps with two segments indicated. The majority of the young, however, are much 
larger. The proboscis and chelifori are well developed ; the two following pairs of limbs 
are quite small, showing few segments and terminating in a long spine or claw. These 
are obviously the palps and ovigers. Behind these are three pairs of well-developed 
]egs, and the body terminates in a rounded sac. The legs bear a few spinous setse and 
more numerous finer ones. 

This species is unquestionably a very close relation of Pallenopsis pilosa, Hoek. 
The spurs on the lateral processes and legs serve to separate it pretty sharply. Dr 
Hoek's species is a deep-water form. 

Nymphon longicoxa (Plate L, figs. 3, 3a). 
Nymphon longicoxa, Hoek, (15), pp. 38-39. 

Specific Characters. — Body and limbs very slender, lateral processes widely separated. 

Palps five-jointed. Proportion of three terminal joints as 4*5, 2, 3. 

Oviger ten -jointed. Denticulate spines numerous, each with three or four lateral 

Legs long, slender, and setose ; second coxa much longer than the other two together ; 

no auxiliary claws. 

Body very slender, and limbs long ; lateral processes very widely separated. 
Segmentation distinct. 

The Cephalon is long, its segment just longer than the proboscis. Just behind its 
centre it is enlarged to form the ventral support of the ovigers ; anteriorly to this it 
forms a very slender neck, and is comparatively little expanded to form the lobes which 
support the proboscis, chelifori, and palps. 

The Ocular tubercle is very short and stout ; it lies just between the first pair of lateral 
processes and the enlargement that carries the ovigers. It bears four distinct eyes. 

The Abdomen is of moderate size, cylindrical, directed upwards, but not articulated 
to the trunk. 

The length of the body is 8 '5 mm. ; of the trunk only, 7 mm. ; and its width is 
3 "5 mm. It is entirely devoid of setse. 

The Proboscis is long, slender, scarcely as long as the cephalic segment ; it is 
cylindrical, slightly swollen in the middle, and rather abruptly rounded distally. 

The Chelifori are long and slender ; the scape is single-jointed, longer than and over- 
reaching the proboscis ; a few small setse are distributed along it, and there is a small 


distal fringe. The chelae are much longer than the scape, and the fingers are longer 
than the palm, which is minutely setose throughout. The dactyli are much curved at 
the tips, and provided with very numerous slender and close-set teeth. 

The Palps arise immediately outside the chelifori. They are long and slender, the 
second joint not quite reaching to the extremity of the proboscis. As usual, the first 
joint is very small, and the second is the longest of the appendage, the proportions 
of the four joints being 7, 4 "5, 2, 3. The second joint bears only a very few setae, 
and these distally ; the third is also scantily setose ; the fourth joint is well provided 
with setae along its ventral margin, and the terminal one is similarly provided, but these 
are smaller and extend to the dorsal surface. 

The Oviger is ten-jointed, the first three joints being short and stout ; the second 
and third are subequal in length. The fourth is long and rather stout ; the fifth is 
more than half as long again, and quite the longest of the appendage. It is greatly 
curved, proximally very slender, but gradually becoming much enlarged distally. It 
bears fine setae along the inner side of the curve, and a row of spinous papillae on the 
opposite side of the enlarged extremity. The sixth joint is curved in the opposite 
direction, and rather more than one-third the length of the preceding. It is covered 
with fine setae on the outer part of the curve. Both these joints are measured across 
the curve. Of the four terminal joints, the first is longest, the other three being 
subequal. The terminal one bears a claw nearly as long as itself; this is armed with a 
dozen slender teeth with fairly wide intervals between them. All the four joints are 
setose dorsally. The denticulate spines are numerous — 15, 9, 8, 9 — and consist of a 
slender tapering and flattened shaft with three rather long lateral teeth on each side, 
and these only rarely arise opposite to one another. The spines are rather worn, 
especially on the two distal joints. 

The Legs are of great tenuity, and attain a length of about 55 mm. Of the three 
coxae, the second is much longer than the other two together, measuring but little less 
than 5 mm. in length. The proportions of the three following are 10, 12*5, 19. The 
tarsus and propodus together measure rather more than 4 mm., the latter being the 
longer joint and carrying a very long slender claw without auxiliaries. The limb is 
setose throughout, the setae being arranged for the most part in rows. A lateral line 
is readily distinguishable. On the first coxae setae are rather scanty ; on the second 
they are not plentiful, except ventrally and distally, where they form a fringe ; on the 
third they are fairly uniformly distributed, and a distal fringe is present. On the 
femur they are sparsely distributed and form a dorsal distal fringe. On the ventral 
surface of this joint there is a row of very small tubercles. On the tibiae the setae 
become much more numerous and longer, especially towards the distal extremity of 
the second, where they approximate to those found on the succeeding joints. On the 
tarsus and propodus they are short and very close set, particularly the ventral row ; 
the dorsal setae are somewhat more spinous. 

The Genital apertures of the male occur on the second coxae of the three posterior 


pairs of legs ; those of the female on all the legs. One male is carrying ova ; these are 
large, and few in number. 

I cannot find any reasonable ground for regarding this species as distinct from 
N. longicoxa, Hoek. The principal differences seem to be the setose nature of the legs, 
and the denticulate spines on the ovigers, which in Dr Hoek's specimens bear seven 
lateral teeth. 

Several specimens were taken in lat. 71° 22' S., long. 16° 34' W., in 1410 fathoms, 
18th March 1904. 

Nymphon compactum (Plate I., figs. 5, 5a). 
Nymphon compactum, Hoek, (15), pj>. 41-43. 

Specific Characters. — Body stout, sparsely hairy ; eyes obsolete. 

Palps five-jointed; proportions of last three, 5, 2*5, 2'5. 

Oviger ten -join ted ; denticulate spines numerous, with four to six lateral teeth. 

Legs long ; auxiliary claws absent. 

Body stout, with the lateral processes short and not widely though very distinctly 
separated. The cephalon is short, much widened anteriorly to form two well-marked 
and divergent lobes for the support of the chelifori. The neck is very distinct, 
narrowest just behind the cephalic lobes ; the space between these and the first pair of 
lateral processes is completely filled by the body-process from which the ovigers arise. 
This body-process is conspicuous from the dorsal aspect. 

The Ocular tubercle lies immediately in front of the first pair of lateral processes ; 
it is stout, of very small elevation, and bears no trace of eyes. 

The Abdomen is pyriform, a little longer than the cephalon ; it is not articulated to 
the trunk. 

The segmentation is not at all prominent, the articulations being immediately 
behind the lateral processes. 

The length of the body is 9\5 mm., of the trunk, 7 mm. ; and its width is 5 mm. 

The Proboscis is directed downwards, cylindrical, and measured ventrally it just 
exceeds 4 mm. 

The Chelifori are well developed. The scape is single-jointed and extends just 
beyond the end of the proboscis. It bears a number of fine long setse arranged in an 
irregular linear manner, and also forming a fairly well-defined distal fringe. The chela is 
long and narrow, the palm being shorter than the slender fingers ; the palm is covered 
with fine setae nearly half way along the immovable finger, and there is a fringe of 
stout setse at the base of the movable one. The fingers are, as usual, much incurved at 
the tips, and provided with very numerous, close-set, slender teeth, not very irregular 
in length. Those of the movable finger are the larger. 

The Palps arise close to the chelifori and are of the normal five joints. The first 



is very short and stout, the proportions of the remainder being 6, 5, 2*5, 2*5. The 
second bears a few scattered setae and a well-marked distal fringe ; the third is rather 
thinly covered with smaller setae, especially distally. On the ventral margin of the 
two terminal joints the setae are stouter than elsewhere and very abundant ; dorsally 
they are rather scanty but much longer. The last joint but one has a well-developed 
distal fringe dorsally. 

The Ovigers are ten -jointed, and arise from stout processes on the ventral surface of the 
neck. The first three joints are very small, and while progressively increasing in length 
diminish in diameter. The three together are a little longer than the fourth ; the propor- 
tions of this and the remaining joints are 8, 10, 7, 3, 2*5, 2, 2'2. The fourth joint is 
curved and stout ; it bears a row of small setae along the greater part of its outer margin, 
and a small distal fringe. The fifth joint is slender proximally, but considerably 
enlarged distally, the transition being gradual and not abrupt ; it is covered, but not 
very thickly, with long slender setae. The following (sixth) joint has its inner border 
curved and is thinly covered with setae finer and shorter than those of the preceding 
joint. Of the four terminal joints, the first two bear a distal fringe only, and the distal 
pair bear a few long setae dorsally in addition. The terminal claw is very nearly as 
long as the joint that bears it, and carries about eleven teeth. Most of them are long, 
curved, and have a considerable interval between them. The denticulate spines are 
numerous — 1 1, 11, 8, 9 — and consist of a tapering flattened shaft with four or five well- 
developed broad teeth on each side ; traces of a sixth may sometimes be detected. The 
second from the base is usually the largest. The spines on the terminal joint are on the 
whole shorter and broader than elsewhere. 

The legs are long, attaining a length of nearly 42 mm. Of the three coxae, the 
second is the longest, but scarcely as long as the other two together ; the proportions of 
the remainder being 9, 11, 10, 35, 2, and 1 for the terminal claw, which is without 
auxiliaries. The limb is setose throughout, but not in any conspicuous manner, the. 
setae being slender and rather long for the most part ; they are arranged in four rows, 
dorsal, ventral, and lateral, and as usual are most conspicuous on the tibiae ; towards the 
end of the second they tend to become spinous. They are small and numerous on 
the tarsus, smaller still on the propodus, where the lateral rows are indistinguishable. 
A lateral line is conspicuous throughout. The distal fringes do not offer any 
special peculiarities, except that of the third coxa, which is very conspicuous 

The Genital apertures of the male are on the second coxae of the two posterior 
pairs of legs ; those of the female are found on all the legs. 

The subject of this description is a fine male, which carries two small spherical 
packets of rather large eggs slung over the proximal part of the fifth joint of each oviger 
by stout threads. 

Three specimens were taken in 1775 fathoms, lat. 62° 10' 5" S., long 41° 20' W. 

There can be no doubt that these specimens are identical with N. compactum, Hoek. 


I have redescribed it on account of small defects in matters of detail in Dr Hoek's 
original description. 

Nymphon capense (Plate L, figs. 2, 2a). 

Specific Characters. — Body stout, with lateral processes distinctly but not widely 

separated. Cephalon with enlarged base. 
Palps five-jointed. Four joints progressively decreasing in size. 
Ovigers ten-jointed ; denticulate spines numerous, each having eight lateral teeth. 
Legs slender ; claw large, without auxiliaries. 

This is a rather small species, with slender limbs. The body is stoutly built, with 
the lateral processes very distinctly but not widely separated. The segmentation is- 
very distinct, and the body is perfectly smooth and transparent. 

The Cephalon is slightly longer than segments two and three together, enlarged 
at its base, dorsal to the origin of the ovigers, and having a well-developed neck, and 
is then widely expanded ; the cephalic lobes are divided nearly to the base by a 
distinct groove. 

The Ocular tubercle is stout, of little elevation, rounded, and carries four well- 
developed eyes ; it lies in front of the first pair of lateral processes. 

The Abdomen is short, not articulated to the trunk, directed upwards, scarcely 
projecting beyond the last pair of lateral processes. 

The length of the body is very nearly 3 mm. ; its width is 1*6 mm. 

The Proboscis is about 1*6 mm. long, measured ventrally. It is rather bottle-shaped, 
stout at the base ; near the middle it is enlarged ; beyond this its diameter is only very 
slightly reduced ; the distal extremity is rounded, and the mouth fairly large. 

The Chelifori are well developed. The single-jointed scape is a little longer than the 
proboscis, and bears a few very small setse and a small distal fringe. The chela is longer 
than the scape, and the fingers longer than the palm and much incurved at the tip. 
The teeth are very closely set, and of three regularly alternating sizes, the space 
between the longest teeth being occupied by two small ones and an intermediate one 
between them. 

The Palps are five-jointed as usual. The first is very small ; the second is the 
longest, and devoid of setse ; the others progressively decrease in length, the proportions 
of the four being 5*5, 5, 3*5, 2*8. The third carries a few small setae, but the fourth is 
most conspicuously setose and has a prominent distal fringe. The setse on the terminal 
joint are short, except distally, and not very abundant. 

The Ovigers are ten-jointed, and rise on a conspicuous body-process just in front of 
the first pair of lateral processes, and very distinct from the dorsum. The first three 
joints are very small, but progressively increase in size, forming a small but distinct 
curve ; the next three form a slight curve in the opposite direction ; these three 
progressively decrease in length, but not by much, the fourth joint of the appendage 


being the longest. All these joints are devoid of setse, except for a distal fringe on the 
sixth. Of the four terminal joints, the difference in size is very small ; the first is the 
longest, and the next two progressively decrease in length, the terminal one being as 
long as the preceding one, and carrying a claw three-quarters of its length. The claw 
bears about fifteen slender teeth rather closely set. The denticulate spines form a 
single row, and are rather long, more gently tapering than usual, especially near the 
apex. They are numerous, the numbers being 15, 12, 11, 12. In uninjured specimens 
there are eight lateral teeth, of which the four basal ones are prominent, the third being 
distinctly the largest ; the others are small and delicate, but the terminal one is a rather 
large blade. All these four joints bear small setse dorsally and a small distal fringe. 

The Legs extend to about 17 mm. They are slender, and only bear a few very small 
setse arranged in rows ; on the ventral surface of the propodus they are most distinct. 
The lateral line is distinct throughout. Of the three coxae, the second is as long as the 
other two together, the proportions of the remaining joints being 7, 8, 107, 3, 3. 
The terminal claw is long and slender, about two-thirds or rather more than the 
length of the propodus. There are no auxiliaries. 

Two males bear eggs. These are large, few in number, and carried in very irregular 
masses. The oviger is but little modified in the male ; the fifth joint is more curved 
and rather dilated distally. 

The Genital apertures of the male are on the second coxse of the two posterior pairs 
of legs ; of the female, on all the legs. 

The body of this species is sufficiently transparent to show some details of its 
anatomy. The " sieve apparatus " is distinctly seen in the proximal half of the proboscis. 
The nerve chain is readily seen as a chain of five ganglia connected by a double cord. 
The ganglia are double and partially fused. The first lies just behind the origin of the 
ovigers and gives off three nerves on each side, the most posterior of which goes to 
the oviger. The other two cannot be traced into any of the appendages. The posterior 
ganglion lies on the line of segmentation of the last segment of the body ; all of them 
give a strong nerve to its appendage. 

In the female from one to four ova can be seen on the second and third coxse and 
the femora. The alimentary canal is distinct in the trunk, but it is not easy to 
determine its limit in the legs. 

A score of specimens, principally females, and some of them in rather a dilapidated 
condition, were taken 8 miles north of Dassen Island, Cape Colony, 18th May 1904, in 
35 fathoms. 

Nymphon articulare (Plate I., figs. 4, 4a). 

Specific Characters. — Body well built, with rather long lateral processes, and these well 

Palp five-jointed ; proportions of the last three 2'5, 15, 1 75. 


Oviger ten-jointed ; denticulate spines not numerous, with six (?) lateral teeth. 
Legs : principal joints subequal, tarsus and propodus subequal, terminal claw with two 
well-developed auxiliaries. 

Body stout, with lateral processes long — the second pair longer than the width of 
the body, and rather widely separated. 

Cephalon not long, constricted to form a well-defined neck, and then expanded to 
form the stout cephalic lobes. 

Ocular tubercle rises between the first pair of lateral processes, abreast of their 
anterior border. It is tall, tapering, and ends in a blunt point where the four eyes 

The segmentation of the trunk is distinct but not very prominent, and the abdomen 
is not separated by an articulation. This part of the body is rather long and slightly 
pyriform. It carries a few very small setae, which also occur as distal fringes of the 
lateral processes. 

The length of body is 3 mm., of the trunk only, 2*25 mm. ; its width is 1*8 mm. 

The Proboscis is almost cylindrical, with a truncated extremity, and its base is 
reduced a little in diameter. Measured ventrally, it is 1*4 mm. long. 

The Chelifori are well developed. The scape is long, 1*3 mm., sparsely covered with 
setae. The chela is strong, the palm being setose ; the fingers are scarcely as long, 
much curved, and beset with numerous closely set teeth of variable length. 

The Palp arises at the side of the proboscis, and comprises the normal five joints ; the 
first is very small and stout; the second is the longest, and its proportionate length with 
the remainder is 4, 2'5, 1*5, 175. The second joint is sparsely covered with rather 
long setae ; on the following joint they are much more numerous, and on the two 
terminals they are more thickly distributed, and chiefly on the ventral surface. 

The Oviger arises ventrally between the first pair of lateral processes. The first 
three joints are very small, the first two being stout and having a very oblique joint 
between them ; the third has the normal oblique joint ; the fourth and fifth are sub- 
equal in length, the former being much the stouter, the sixth being a little more than 
half their length. Very few small setae are to be found on this part of the appendage ; 
they are, however, rather more numerous on the sixth joint. The four terminal joints 
are very nearly subequal, setose dorsally, and each has a row of a few denticulate spines, 
6, 6, 5, 7. The terminal claw is rather long, with about nine teeth not very closely set 
These and the denticulate spines are rather worn. The spines are curved forwards, 
probably due to pressure. There are four well-developed teeth on each side, and 
probably two more delicate ones. 

The Legs attain a length of about 9 mm. Of the three coxae, the second is the 
longest, but scarcely as long as the other two together ; the proportions of the other 
joints are 4, 4, 4, 2, 2. The terminal claw is about half the length of the propodus, and 
has two well-developed auxiliaries. The entire limb is covered with rather fine setae ; 


they are not very abundant, and the linear arrangement is not distinct ; a distal fringe 
is conspicuous ventrally on the third coxa, more so dorsally on the femur ; on the 
second tibia it is ventral, and the setae are distinctly spinous but few in number. Setae 
are scarce and very small on the tarsus and propodus, and on these joints it is usual to 
find a ventral row which is to some extent at least characteristic. In this species this 
row consists of very inconspicuous setae, and placed at comparatively large intervals. 

Three specimens were found among a large number of Chastonymphon orcadense. 
They are all adult females, and attention was attracted to them by the peculiar enlarge- 
ment of the femora. These joints are not enlarged throughout, as in most species, but 
considerably bellied for about two-thirds of their length. 

Ch&tonymphon brevicaudatum. 

Nymphon brevicaudatum, Miers, (20), pp. 200-214. 
Nymphon horridum, Bohm, (1), p. 172. 
Nymphon brevicaiidatum, Hoek, (15), pp. 49-52. 

Three specimens of this species were taken in Scotia Bay, South Orkneys. They 
were found amongst an immense number of Chsetouymphon orcadense captured during 
the autumn and winter of 1903 inside the 15-fathom line. The specimens are: 
an adult female, an ovigerous male, and a small one of which the sex is uncertain. All 
of them show the setae of the body arranged in the stellate manner described and 
figured by Dr Hoek ; but that author distinctly states that the setae are not placed in 
regular rows on the legs. In these specimens this is a striking feature ; the setae are 
long and coarse, arranged in five rows, two dorsal, two lateral (and these are the longest), 
and a single ventral row, where they are shortest but most numerous. This arrangement 
is most noticeable on the tibiae. Smaller and finer setae are also irregularly distributed 
over the appendage, but more particularly on the ventral surface of the coxae, especially 
the third, and the femora. The tarsus and propodus are much more slender than the 
rest of the appendage, the latter being a little the longer, and here the setae are small 
and there is a well-developed ventral row ; dorsally and laterally the linear arrange- 
ment is indistinct. The terminal claw is accompanied by two well-developed 

The ova carried by the male are not very numerous, but large. 

Chzetomymphon mendosum. 
Chxtoiiymphon mendosum. Hodgson, (13), pp. 30-32. 

Specific Characters. — Body robust and tapering ; articulation imperfect; lateral processes 
not widely separated, and with stout spines distally and dorsally ; no fine setae 

Ocular tubercle short and stout. 

Palp five-jointed ; proportions of last three 5*5, 1*5, 1 "G. 


Oviger ten-jointed ; denticulate spines few, with four teeth on each side, two of them 

Legs with five rows of spinous setae, without enlarged bases ; terminal claw with two 

small auxiliaries. 

I have fully described this species in the Eeport quoted above. Its occurrence on 
the opposite side of the Polar area is a matter of great interest. No less than thirteen 
specimens of both sexes were found among an immense number of Chsetonymphon 
orcadense from Scotia Bay, taken in depths of less than 15 fathoms. As in the case 
of those taken by the Discovery, the body is quite smooth, and while there is a fair 
amount of variation in the setose covering of the appendages, its arrangement is 
characteristic. The lateral processes carry two stout spines dorsally and distally ; but 
the number is not confined to two. Two other smaller ones may be found outside the 
principal ones. The spinous armature of the coxae varies more, and while the setae on 
the femur are not so regular in their distribution, the characteristic five rows of strong 
setse are conspicuous on the tibiae, the two dorsal rows having the smaller setae. One 
specimen — the largest, and an adult female — has its setose covering the least well 

The Ocular tubercle of all these specimens is rather flattened and tapering from a 
broad base. 

The Denticulate spines, which in the type specimens were rather worn, are here well 
preserved. The large female already alluded to has the same number as the type ; the 
males may have fewer. 

The shaft tapers, and may have as many as five lateral teeth, the fifth being little 
more than a trace. 

Chwtonymphon orcadense (Plate II., figs. 2, 2a). 

Specific Characters. — Body robust, with lateral processes close together, setose. 
Palps five-jointed; proportions of three terminal joints 4, 2*5, 2*5. 
Oviger ten-jointed ; denticulate spines rather numerous, with 5-7 lateral teeth. 
Legs coarsely setose, with terminal claw and well-developed auxiliaries. 

Body robust, with lateral processes of unequal length, giving it an ovoid form ; they 
are distinctly though not widely separated, variable, apparently depending on age. 
Segmentation is very distinct. 

The Cephalon is not long, constricted near the middle to form a distinct neck, and 
then widely expanded, forming two prominent divergent lobes. In length it is about 
two-thirds of the first trunk segment. 

The Ocular tubercle is rather stout, not quite clear of the first pair of lateral pro- 
cesses. It is tall, rounded at the extremity, where there are four well-developed eyes. 

The Abdomen is not articulated to the trunk, very distinctly pyriform in shape, and 
not quite so long as the cephalon. 


The entire animal is covered with setae ; on the body they are quite small. The 
length of the body is 8 mm., its width 5 '5 mm. 

The Proboscis is not half the length of the body, cylindrical, with a constricted 
base, and truncate at the extremity. It is finely setose throughout. 

The Chelifori are strongly developed. The scape is a single joint longer than the 
proboscis, richly supplied with setse, which are most conspicuous on the inner side, 
some of the dorsal ones being very powerful ; the distal fringe is strongly developed, 
the setae being of varied size. The chelae are as long as the scape, the palm fairly 
stout and covered completely with short setae, which extend well on to the immovable 
finger. The two fingers are slender, much curved at the tips, and furnished with a 
large number of closely set teeth, which are fairly regular in size. 

The Palps are five-jointed, and rise below the chelifori, at the side of the proboscis. 
The first joint is very small ; the second is the longest ; the proportion of that to the 
remainder is 5, 4, 2*5, 2 "5. The entire limb is setose, setae being fairly plentiful 
on the second joint and becoming more numerous to the distal extremity of the 
appendage. They are thickest on the ventral surface of the two terminal joints ; the 
third and fourth have well-developed distal fringes. 

The Oviger is ten-jointed, and rises ventro-laterally just in front of the first pair of 
lateral processes. The first three joints are very small and stout, the second and third 
being subequal, the latter having, as usual, a very oblique termination ; the fourth is 
slightly curved, about as long as the first three together, and stouter distally than 
proximally ; the fifth is fully half as long again, more strongly curved, enlarged 
distally, and having a curious constriction about a quarter of its length ; the sixth is 
about one-third the length of the fifth, and has a well-developed setose tubercle at its 
extremity ; the seventh joint is articulated at the outer angle of the sixth, but this is 
not always easy to observe. The four terminal joints differ but little in size ; the first 
three progressively decrease in length, and the fourth is as long as the second. 
The terminal claw carries upwards of a dozen slender teeth. The denticulate spines 
are in a single row of 12, 8, 7, 10 on the four joints. Each spine is a stout blade 
with from five to seven lateral teeth ; the middle ones are the largest, and the 
distal ones only exist as a mere trace. The entire appendage is setose ; the setae are 
small ; the fourth joint carries them on the outer side of its curve, the fifth also, but 
distally they are much more evenly distributed ; the sixth is completely covered, and 
the four terminal joints are provided dorsally with numerous setae, longer than else- 
where. The oviger of the female presents considerable differences from the above. 
The limb is much more slender ; the fourth and fifth joints are large, slightly curved, 
stoutest distally, and subequal in length. The setose tubercle at the extremity of the 
sixth joint is not noticeable. The number of denticulate spines on the various joints 
is not quite constant, nor are the teeth on the terminal claw. The setose covering is 
not so well developed. 

The Legs attain a length of 32 mm. Of the three coxae, the second is about as long 


as the other two together ; all are fairly uniformly covered with short setae ; the distal 
fringe of the third is, however, ventral and composed of numerous long slender setae, and 
these are most conspicuously developed on the two posterior pairs of legs of the male 
— obviously a sexual character. The three following joints differ but little, their pro- 
portions being 6'5, 7, 7'25 ; the tarsus and propodus being as 3 to 2 - 5. The terminal 
claw is rather slender, and is accompanied by two auxiliaries less than half its size. 
The limb is thickly covered with setae of no great size, but variable in length ; they 
are longer and stronger laterally than elsewhere. This is not so conspicuous on the 
femur as on the tibiae ; they are largest on the second tibia, though the general covering 
of this joint is much finer than on the preceding joints. On the tarsus and propodus 
the setae are small, uniform in size, and thickly set. The distal fringe of the femur is 
dorsal and composed of long stiff setae ; that of the first tibia is ventral and rather more 
spinous ; that of the second tibia is also ventral and composed of strong spines. 

The Genital apertures of the male are only to be found on the second coxae of the 
two posterior pairs of legs ; those of the female occur on every leg. 

Of the enormous number of specimens taken in Scotia Bay, but a few males carry 
eggs. These are rather large, 0'7 mm., and between thirty and forty in a packet. This 
is slung over the fifth joint of the oviger by a stout thread near its proximal end. 
Another specimen carries a very large mass of young. These show the chelae well 
developed, as are also the first two pairs of legs ; the third pair are conspicuous rudi- 
ments, and the fourth pair can be seen as a swelling on either side of the abdomen. 
Rudiments of the palps are visible, and the ocular tubercle is present. Scotia Bay, 
about 10 fathoms. Burdwood Bank, south of the Falkland Islands, one specimen, 
immature, 56 fathoms. 

Chsetonymphon assimile (Plate I., figs. 1, la). 

Specific Characters. — Body robust, with lateral processes very distinctly separated ; 

devoid of setae but for the distal fringes. 
Palps five-jointed ; proportions of last three 7 '3, 4, 4. 

Ovigers ten-jointed ; denticulate spines not numerous, with five (?) lateral teeth. 
Legs with long terminal claw and very minute auxiliaries. 

The Body is stoutly built, with the lateral processes distinctly but not widely 
separated, and they carry a few stout setae as a distal fringe ; otherwise the body is 
quite smooth. Segmentation is distinct but not prominent, and the abdomen, which 
is pyriform and directed upwards to clear the posterior lateral processes, is not articu- 
lated to the trunk ; it only bears two small setae near its extremity, sometimes more. 

The Cephalon is short, constricted to form a distinct neck, and then expanded to 
form two diveroent lobes. 

The Ocular tubercle lies between the neck and the first pair of lateral processes as 
a tall cylindrical structure, having at its rounded summit four well-developed eyes. 



The length of the body is very nearly 6 mm., and its width 3 mm. ; the abdomen takes 
up 1*2 mm. 

The Proboscis is cylindrical, and is thinly covered with small setae ; the extremity 
is truncated. 

The Chelifori are strongly developed. The scape is single-jointed and 2*3 mm. long ; 
it is covered with small setae and a row of three or four spinous ones along the mid-dorsal 
surface, and a distal fringe of the same kind. The chelae are about as long as the scape, 
the dactyli being longer than the palm, which is covered with very short setae well on 
to the base of the immovable finger. Both fingers are much curved at the tips and 
bent on their inner margins, with numerous closely-set, slender teeth not of uniform size. 

The Palps arise at the side of the proboscis, and comprise the normal five joints. 
The first is small and stout ; the second is the longest, but only by a very little, the 
proportions between it and the remaining joints being 8, 7 "3, 4, 4. The second joint 
bears a few setae scattered along it, and two or three distinctly spinous ones, and a 
distal fringe of stout setae. The third joint is more plentifully supplied with setae, 
more uniformly distributed ; the distal fringe is well developed ventrally. The two 
following joints are still more richly supplied on the ventral surface, but not to the 
same extent dorsally. 

The Ovigers arise in the angle between the first lateral process and the cephalon, 
the body-process being distinctly visible dorsally. Of the ten joints, the first two are 
very small and stout ; the third is as long as the two together. The fourth and fifth 
are almost subequal, the fifth being a very little the longer ; the former is stout, the 
latter more slender and covered on its outer margin with short stiff setae ; the sixth is 
comparatively long, just exceeding two-thirds the length of the fifth, and is similarly 
setose. The four terminal joints differ but little in size ; the third is perhaps the 
smallest ; the dorsal surface of all is well provided with rather long setae, and the 
terminal claw carries about a dozen teeth. The denticulate spines are not numerous 
(9, 8, 6, 9). They are much worn, as are also the teeth on the terminal claw, but 
exhibit a tapering shaft with three strongly developed lateral teeth and probably two 
more delicate ones. 

The Legs attain a length of 20 mm. Of the three coxae, the second is not so long 
as the other two together ; the proportions of the remaining joints are as 8, 11, 10, 4"5, 3. 
The terminal claw is long and slender, and the auxiliaries are extremely minute. 
The entire limb is setose. The setae on the coxae are scanty dorsally, being limited to 
a poorly developed fringe on the first, and a lateral row on the second ; but the ventral 
surface of the second and more particularly the third are covered with short stiff setae 
and distinct distal fringes. The femur is irregularly clothed with rather long setae 
dorsally, and a prominent distal fringe ; ventrally the setae are quite small. A similar 
arrangement holds good on the two tibiae, but there is in addition a lateral row of stout 
spinous setae situated at rather long intervals. The distal fringe of the first tibia is 
inconspicuous but for one stout spine ventro-laterally, and on the second tibia it is 


composed ventrally of stout spines. On the tarsus and propodus the setae are small 
and inconspicuous ; there is a row ventrally along the two joints in which the setae are 
more regular than elsewhere, but there is nothing very distinctive about them. 

The specimen described above is an adult female, and the genital apertures occur, 
as usual, on all the legs. Those of the male occur only on the two posterior pairs of legs. 
The male differs further in being more coarsely setose, though the arrangement of the 
setae is identical. The fifth joint of the oviger is abruptly expanded to more than twice 
its normal diameter, and the sixth joint is also expanded considerably, but is reduced to 
something like normal diameter distally. The denticulate spines are a little more 
numerous than in the female, being 11, 7, 6, 8. 

A number of specimens were taken in Scotia Bay, in less than 10 fathoms, during the 
stay of the Scotia. 

Pentanymphon antarcticum. 

Pentanyrnphon antarcticum, Hodgson, (10), pp. 458-462; (11), p. 35; (13), pp. 36-39. 
„ „ Cole, (6), p. 105 ; Bouvier, (3), p. 4. 

Only a single specimen, an adult female, was taken by the expedition, and was found 
among a large number of Chsetonymphon orcadense from Scotia Bay, in 10 fathoms. 
It is more robust than the average specimens from M'Murdo Sound ; the scabrous 
nature of the body is more distinct, but this is not readily noticeable. Its neck, 
though a little shorter in proportion to the type specimens, does not appear to be more 
than an individual peculiarity. The joints of the legs have the same proportions as the 
types, but the tarsus and propodus vary on the different limbs more than usual for a 
single individual. The denticulate spines of the ovigers are much worn, and quite 
useless for specific discrimination. The genital apertures occur on the second coxa of 
all the legs. 

Scotia Bay, 10 fathoms. 

Ammothea communis (Plate II., figs. 1, la). 
Ammothea communis, Bouvier, (3), p. 6. 

Specific Characters. — Body discoid, with lateral processes close together ; they and the 

first coxae armed dorsally with two stout spurs. 
Palps eight- jointed, last three progressively increasing in length. 
Oviger ten -jointed, small ; few denticulate spines, no claw. 
Legs stout, uniformly setose ; terminal claw stout, with two well-developed auxiliaries. 

Four strong spines on propodus. 

Body discoid, the lateral processes being close together ; these are armed with two 
stout spurs dorsally and distally, and similar but stouter ones occur on the first coxae. 
Of these the posterior one is always the largest. 


The Cephalon is stout, and also provided anteriorly with a short but stout spur on 
each side, and between these is the stout Ocular tubercle. This is short, blunt, and 
rounded, but bears a small pointed tubercle at the apex. Of the four well-developed 
eyes, the anterior pair are the larger. 

The Abdomen is rather long and cylindrical, directed upwards, but not so much so 
as in Professor Bouvier's specimens. Very small setae occur on the abdomen and all 
the spurs of the cephalon and lateral processes. 

The length of the body is 17 mm., and its breadth 1*3. 

The Proboscis is large, pyriform, rather more than half the length of the body. 

The Chelifori are as is usual with the members of this genus ; the scape is short, 
slightly curved, with but few minute setae, and those distally. The chela is reduced to 
a mere knob, with but traces of the dactyli. 

The Palps are eight-jointed and rise laterally to the proboscis. The first joint is 
short and stout ; the second and fourth are subequal and about three times as long, the 
latter bearing a few setae ventrally. The third joint is quite small, with two or three 
setae dorsally. The four terminal joints are all small, and differ but little in size ; 
however, the last three progressively increase in length, the terminal one being distinctly 
the longest and rather irregular in shape ; it is richly supplied with short stiff setae 
ventrally and distally, to a less extent dorsally. The other joints are enlarged a little 
ventrally, and carry a tuft of short stiff setse. 

The Oviger is ten-jointed, short, with all the joints small. The first is short and 
stout; the second is much longer and more slender; the third, shorter than the pre- 
ceding one ; the fourth and fifth are subequal, slightly curved, and the longest of the 
appendage, the proportions of the third to these being 3, 4, 4. The sixth is small, 
the seventh a little longer and more slender ; the remainder graduate to a very small 
terminal joint. Very few seta? are to be found on the entire appendage, and the 
denticulate spines are also in very small numbers (l, 2, 1, 2). They consist of a 
slightly curved shaft with the flattened blade cut into nine very closely set teeth. On 
the sixth joint there are two spines, and another on the seventh, which look as if they 
were much-worn specimens of the denticulate spines. This would increase the formula 
to 2, 2, 2, 1, 2. 

The Legs attain a length of 6 mm. The first coxa is stout, with two dorsal spines, 
as before stated ; the second is longer, slender, but enlarged distally, not as long as the 
other two together. The proportions of the three following joints are as 6, 6, 5. The 
tarsus is very short ; the propodus is curved and bears a stout terminal claw and two 
very well-developed auxiliaries. The entire limb is covered with short and rather stiff 
setae. The tarsus sometimes bears a stout spine ventrally. The propodus has four 
stout spines ventrally and proximally, the remainder of the ventral surface being 
occupied by a band of stout setae. The difference between these and the four proximal 
spines is not so sharply marked as in some species. The femora bear a stout tubercle 
dorsally and distally. 


The Genital apertures are on a prominent outgrowth at the distal extremity of 
the second coxae of the two posterior pairs of legs. The eggs are numerous, and two 
roughly spherical packets may be found on each oviger. 

Two specimens, both ovigerous males, were taken in Scotia Bay, at a depth of 
10 fathoms. 

Leionymphon grand e. 

Ammothea grandis, Pfeffer, (24), pp. 43-45. 
Oolossendeis (?) Chareoti, Bouvier, (2), pp. 295-296. 
Leionljmphon grande, Hodgson, (13), pp. 41-43. 

This species Avas first described by Dr Pfeffer as coming from South Georgia. 
Professor Bouvier has recorded it more recently from Carthage Bay, where it was 
taken by the French Antarctic Expedition. A single adult female and several 
immature specimens were captured by the Discovery off Coulman Island in the Ross 
Sea. In the Report on the Discovery Pycnogonida I have described the species at 
some length, and transferred it to a genus founded by Professor Mobius (22) for an 
immature species taken off Bouvet Island. 

A single specimen only was taken by the Scotch Expedition in Scotia Bay, South 
Orkneys, 14 fathoms, 26th March 1903. It is a male, not quite adult, retaining the 
chelate condition of the chelifori ; and the genital apertures are not yet developed. In 
point of size it is a good deal smaller than the adults, but otherwise does not exhibit 
any important differences. The length of the body is 11 mm., its width 9 - 5 mm. 
The proboscis, which tapers very slightly towards the extremity, is 12 mm. long. 
The legs have a length of only 42 mm., the proportions of the three principal joints 
being 9*5, 9, 12. There are only three stout spines proximally on the ventral side of 
the propodus, and the distribution of the short stiff setae over the legs is quite uniform. 
The ocular tubercle ends in a cone above the eyes. 

These are the only differences to be found between this specimen and the adult 
taken by the Discovery. 

The oviger, however, presents important sexual features, and though this specimen 
is not mature the appendage may be described in detail. The first joint is very small ; 
the second is longer and stout ; the third is more slender, and has, as usual, a very 
oblique termination ; if measured to the extremity of this it is about as long as the 
preceding. Their outer margins are rather thickly covered with short setse. The 
three following joints form a curvature in the opposite direction to the first three. 
Their proportions are about as 5, 6, 3*5, and they are covered with short setae, but 
more particularly so on the outer side. The precise proportions of the remaining 
joints .cannot be given, owing to their relations one to the other. The seventh 
joint is shorter than the preceding, and articulated to it at something like a right 
angle. Near its distal extremity it bears a tuft of setse. The eighth joint is richly 
setose, and also articulated to the seventh at a considerable angle ; the two terminals 


taper to a blunt point, are devoid of setse, and do not appear to have attained their full 

Leionymphon Claud (Plate II.. figs. 1, la). 

Ammothea Clausii, Pffeffer, (24), pp. 43-44. 
Leionymphon Clausi, Hodgson, (13), p. 40. 

Specific Characters. — Body with lateral processes not widely separated but divergent. 

Abdomen tall, erect, immediately behind posterior trunk segment. 
Palps nine-jointed, the last five differing but little in size. 
Ovigers ten-jointed, without terminal claw, and differing in the two sexes. 
Legs with strong terminal claw and auxiliaries ; about five stout spines ventrally and 

proximally on the propodus. 

The Body is robust, with the lateral processes rather long, distinctly but not widely 
separated, divergent. Segmentation rendered conspicuous by three prominent trans- 
verse ridges, which are rounded and not produced into a median point. 

The Cephalon is but very little expanded. Its anterior margin is straight, and it is 
about two-thirds the length of the anterior segment. At its antero-lateral angle it 
bears a distinct tubercle, and there are two more similar ones on each side of the lateral 
processes. Here the posterior one of each pair is the largest. 

The Ocular tubercle lies slightly in advance of the middle line of the cephalon ; it 
is tall and cylindrical, bearing four well-developed eyes, above which it terminates in an 
elongated cone. 

The Abdomen is directed straight upwards, and lies so close to the posterior segment 
that the line passes over its base. It is cylindrical, tapering distally to a blunt point. 
The length of the body, measured to the extremity of the posterior lateral processes, is 
barely 6 mm. ; its width is 4 mm. 

The Proboscis is pyriform, quite smooth, and fully 5 mm. in length. Movably 
articulated to the body, it is carried directly downwards at a considerable angle. 

The Chelifori are rudimentary ; they arise from the anterior margin of the cephalon 
above the proboscis, and are curved, having a few setse distally. The chelse are irregularly 
shaped knobs inclined downwards, with small tubercles to represent the dactyli. 

The Palps arise just below the chelifori at the sides of the proboscis, and are nine- 
jointed. The first joint is short and stout ; the proportions of the three following are 
as 6, 1, 4'5 ; the remaining five differ but little in size, and, retaining the proper numbers 
of each joint, their sequence in point of length is as follows : 5, 7, 9, 6, 8. The eDtire 
appendage is covered with very small setse, and these are specially numerous on the 
ventral surface of the five terminal joints. 

The Ovigers are ten-jointed and arise ventro- laterally from a small body-process on 
the neck, and therefore immediately in front of the first pair of lateral processes. 


Between the two sexes they exhibit considerable differences. The type specimens are 
males, and the specimen now under examination is an adult female. 

All the joints are small ; the first is short and stout ; the second, fourth, and fifth are 
subequal, and, by a little, the longest joints of the appendage ; the third is about two- 
thirds the size of these ; the sixth is a little shorter than the preceding ; the seventh 
and eighth are a little shorter still and subequal ; the ninth is the shortest of all, except 
the first ; and the tenth is slender, rather tapering, and half as long again as the ninth. 
Setse are very scarce throughout, except on the sixth joint, where they are fairly 
numerous though minute. The last four joints bear numerous denticulate spines, which 
are not, however, arranged in a single row. These are unfortunately much worn, but 
show a stout, usually curved shaft, bent, with some eight or nine very closely set teeth 
on each side. There is no terminal claw. 

The Legs extend to a length of about 27 mm. Of the three coxae, the first is the 
shortest, the second is rather more than twice as long, and the third is intermediate 
between the two. The proportions of the three following joints are as 6, 5, 7, but they 
are not strictly preserved on all the legs. The tarsus is very small, covered with stiff 
setse, which become distinctly spinous as a ventral distal fringe. The propodus is rather 
curved ; ventrally and proximally there is a row of very stout spines, five in number. The 
rest of the joint is thickly covered with short stiff setae, especially ventrally ; dorsally 
there are a few longer ones. The heel does not project much and is fringed with stiff 
setse. The terminal claw is long and stout, with two large auxiliaries more than half as 
long ; the three rise from a common membranous investment. 

The limb is covered fairly plentifully with short stiff setse, and there are rows on the 
principal joints of stout spinous setse dorsally and laterally ; these are, however, not 
closely set. The distal fringes are not conspicuous. 

The Genital apertures are found on the second coxse of all the legs. 

A single specimen was found among an enormous number of Chsetonpmphon 
orcadense taken in Scotia Bay, at 9 fathoms. 

Through the courtesy of Dr Pfeffer of Hamburg, I have been able to compare 
this specimen with the types of his Ammothea Clausi. Though Dr Pfeffer's 
specimens are both ovigerous males, there cannot be any question as to the identity of 
this specimen with that species. The ova are small and numerous, massed together in 
two spherical packets on each oviger. 

Decolopoda australis (Plate III., figs. 2, 2a, 26, 2c). 

Decolopoda australis, Eights, (8), pp. 203-206 ; Stebbing, (29), pp. 185-189 ; Cole, (6), pp. 405-415 ; 

Loman, (19), p. 722. 
Decolopoda australis, Hodgson, (11), pp. 36-42; (12), pp. 254-256; Bouvier, (3), pp. 1-5. 

This fine species was discovered in some numbers along the shores of the South 
Shetland Islands, and a very full description was published by Dr Eights in the first 


volume of the Boston Journal of Natural History in 1837. The description is clear and 
accurate ; if the plate which accompanies the paper is not as good as it might be, it is 
readily recognisable, and quite as good as many produced at the present day. It is 
more than probable that the title of the paper has contributed largely to its having 
been overlooked for so long ; but that does not justify the reception allotted to it when 
discovered. It would appear that a Pycnogonid with five pairs of legs was not to be 
tolerated, and two well-known zoologists have failed to recognise such a possibility. 
However, the species has been rediscovered, and another totally distinct and not closely 
related species, also with five pairs of legs, has been discovered in the distant south. The 
modern discovery which has led to a proper appreciation of Eights' species has been 
achieved by the recent Antarctic expeditions, three of which secured a number of 
Eights' species from the South Orkneys, in latitude 61° S. The discovery of such 
an interesting form was deemed of sufficient importance to warrant its prompt publica- 
tion, and it was not till that was undertaken that the significance of Eights' work came 
to light. In a communication made to the Royal Physical Society of Edinburgh (11), 
Eights' original description has been republished verbatim, together with a more 
modern one. It is a matter of opinion which is the better of the two ; the plates, how- 
ever, are much superior, and give a very accurate idea of the animal. 

The more recent description is republished here, with some verbal alterations, and 
some further information is added in a final paragraph. 

Decolopoda australis, Eights. 

Specific Characters. — Body stout, discoid, with only occasional traces of segmentation, 

with a group of three or four very small spines on the dorsal surface of the lateral 

processes, which are quite close together. 
Proboscis large, bent downwards just beyond the middle of its length, with minute 

spines dorsally. 
Palps : third joint considerably the longest ; the three terminal joints subequal, and 

shorter than the seventh. 
Legs smooth ; setse restricted to spinous distal fringes. 

In comparing the South Orkney specimens with Eights' description, one is struck 
by the accuracy of that naturalist ; but according to modern requirements some small 
points have been overlooked. 

Eights describes his specimens as being a bright scarlet, and the body and coxae of 
the figure have been so coloured. The South Orkney specimens, after being in spirit 
for more than a year, do not show any trace of such a colour. Some of the specimens are 
of a very light straw colour, without any trace of pigment, except in one or two cases 
where a little is distributed at the extremity of the proboscis. Other specimens are of 
a rich olive-brown colour, which is considerably darker, almost black, on the proboscis, 
mandibles, and palps. In one specimen the legs are equally darkly pigmented. The 


colour notes taken at the time, and which have been forwarded to me, show that some 
of the specimens were bright scarlet, as Eights described them ; others are a very dark 
red, and in the latter case the proboscis is almost black. The scarlet colour appears to 
be uniformly distributed over the body and limbs, the proboscis and adjacent parts 
being darker than the rest. One specimen shows a distinct segmentation, two others 
show it very indistinctly, and the remainder not at all. 

The Cephalon is short ; a distinct neck separates it from the rest of the trunk. It is 
expanded to completely fill the interval between the first pair of lateral processes. 

The Ocular tubercle lies in the middle of the cephalon, and is short and stout ; it 
bears four well-developed eyes below the terminal cone. 

The Proboscis is not so long as the body, but longer than the trunk. It is much 
swollen just beyond the middle, where it bends downwards at a considerable angle. 
The mouth is small. Along the middle line of the proboscis is a narrow band of small 
spines, which can hardly be said to have a regular arrangement. More laterally are 
two or three rows of spines, not always well defined ; the inner one comprises several 
spines, but the outer one only a very few. On the whole, the lateral spines are larger 
than those of the median band. One or two spines may occur ventrally just behind 
the bend. 

The Chelifori are well developed and the scape is two-jointed. They arise from the 
wide extremity of the cephalon laterally to the proboscis. The first joint is long, 
reaching almost to the beginning of the median enlargement of that organ. The second 
joint is very small, and constitutes the angle in the direction of the appendage ; seen 
laterally, it is practically triangular in shape, the apex downwards and its dorsal margin 
sinuous. The chela is articulated to the ventral angle, and lies close underneath the first 
joint. The palm is very small ; both fingers are slender and much curved, like a pair of 
callipers ; the tips overlap, and there is no trace of teeth. There are no setae on the 

The ten-jointed Palps rise ventrally close against the proboscis. It is open to 
question if the first is a true joint or merely a body-process ; it is, however, constricted at 
the base. The first two joints are very short and wide — annular, in fact ; the third is the 
longest in the appendage, and in proportion to the three following is as 6, 1, 4, 2. The 
following joints are all small and differ but little in length. The seventh is perhaps 
the longest, the eighth the shortest ; the two terminals are subequal and longer, but 
the difference is scarcely noticeable. The fifth joint is the stoutest of them all, and 
slightly constricted about two-thirds of its length. In none of the specimens are the 
palps setaceous, and this is the only discrepancy between Eights' description and the 
South Orkney examples. 

The Ovigers rise from a body-process immediately behind the palps, close to the 

middle line of the ventral surface. This body-process is quite as large ventrally as any 

of the first three joints, which are small, and might easily be mistaken for one. The 

proportions of the three following joints are as 9 "5, 5, 10. The four terminal joints 



are subequal in length, and bent on each other to form a loop which is so character- 
istic of the genus Colossendeis. Each joint is furnished with four rows of non-denticu- 
late spines, though on the terminal joint these spines are less regularly arranged, and 
a fifth row may be distinguished. The spines vary a good deal in form and size ; the 
larger ones are trenchant blades, bent rather close against the surface from which they 
spring. The terminal claw is long and slender. These spines, as well as the terminal 
claw, are much worn in all the specimens available, though of course to a varying 
extent, and I would not assert that I had seen a perfect specimen. It does not appear 
to me improbable that they should prove to terminate in a flattened blade with a 
denticulate margin. 

The Leg extends to a length of about 83 mm. They differ in length, but without 
measurement the difference is hardly noticeable. The fifth leg is the smallest, the 
first comes next, and the fourth is intermediate between the third and fifth ; the second 
and third are the largest and subequal. Of the three coxse, which are short, the first 
is the shortest, and the second the longest, the differences being small. The first is 
marked dorsally and ventrally by a faint groove passing along the greater part of the 
joint, and rendered more conspicuous by a change of colour. Its distal extremity 
is marked by a minute spine. The other coxa3 exhibit distinctly the " lateral line," 
which is continued to the extremity of the limb. The proportions of the other joints 
and claw are as 18, 19, 20'5, 8'5, 7, 5*5. There is a certain amount of variation, but it 
is confined to narrow limits. Except for the spinous distal fringes, the limbs are 
perfectly smooth. The distal fringe of the femur consists of one mid-dorsal spine and 
two lateral ; these are quite small and inconspicuous ; that of the first tibia is generally 
deficient ; those of the second tibia and tarsus comprise four long spines, each pair being 
situated ventro-laterally ; a mid-ventral spine occurs occasionally. A pair of long 
spines occurs ventrally on the propodus, one on each side of the claw. 

The Genital apertures occur on the second coxa of all the legs ; those of the male 
are small and distal, those of the female are larger and median. 

Several specimens of this fine species were taken on various dates in Scotia Bay, 
South Orkneys, in 9 or 10 fathoms. Two are quite immature, but show no traces of seg- 
mentation, and in all particulars, except the non-development of the genital apertures, 
resemble the mature specimens. Two specimens in the collection were marked as 
bearing ova, but examination shows that these "ova" were bunches of contracted 
hydroid polyps, and were held by the mandibles, but close against the ovigers. 

Colossendeis orcadense (Plate II.. fig. 3). 

Specific Characters. — Body well built, with lateral processes widely separated. 
Proboscis considerably longer than the trunk, and enlarged from near the middle 

Palps ten-jointed; proportions of last three joints 1'8, 2*5, 2'a. 
Ovigers ten-jointed, the last four joints with five rows of denticulate spines. 


Ocular tubercle stout, conical, without eyes. 

Legs slender, covered with rows of very minute setae ; terminal claw long. 

The Body is fairly robust, without any trace of segmentation, and the lateral 
processes are widely separated. The cephalon is small, and at its posterior border, not 
quite clear of the first pair of lateral processes, lies the ocular tubercle. This is a very 
short but stout and pointed process, without any trace of eyes. The abdomen is slender 
and articulated to the trunk. 

The Proboscis is considerably longer than the body, and is bottle-shaped — that is, 
it is narrow at the base and for about a third of its length, when it becomes consider- 
ably enlarged. At about its distal third it again becomes slightly constricted, and 
terminates in three lobes corresponding to the sides of the mouth. It is irregularly 
covered with minute setae, and movably articulated to the trunk. 

The length of the proboscis is 20 mm. ; trunk, 13 mm. ; abdomen, 2 mm. ; width 
of trunk, 7 mm. 

The Palps arise ventro-laterally at the extreme end of the trunk. The first joint, 
though small, is unusually large, and twice the size of the second, which is annular. The 
proportions of the other joints are 11, 1*3, 7, 4, 5*75, 1"8, 2 - 5, 2'5. 

With the exception of the first two joints, the entire appendage is setose throughout. 
On the third joint they are scarce and hardly noticeable, but become more numerous on 
succeeding joints. From the sixth they are most abundant, always small, and thickest 
on the ventral aspect of the appendage. On the fifth joint, at about three-quarters of 
its length, there is a distinct swelling externally and a faint constriction, a peculiarity 
which seems to prevail throughout the genus. 

The Ovigers lie immediately behind the palps, and the processes on which they arise 
are quite close to the middle line and seem to have pushed the palps off the body. 
Each might quite justifiably be taken for an eleventh joint ; the first three, properly 
so called, are quite small and approximately subequal. The proportions of the next 
three are 17, 7'5, 15 '2. The four terminal joints present no exceptional peculiarities 
in general appearance, but the denticulate spines are arranged in five rows altogether — 
a single row of large spines separated by an interval from a second which contains 
nearly double the number of smaller spines, another interval, and then three more rows 
not so regular as the other two. In the ninth joint these latter rows are very irregular. 
The spines themselves are of the characteristic type — a more or less cylindrical base 
bearing a flattened blade, the two being subequal in length. In the present specimen 
they are so much worn that no details as to their margins can be observed. 

The Legs are slender and are 110 mm. in length. The three coxae are small and 
subequal, the first being very closely articulated to its lateral process, the proportions 
of the remaining joints being 32, 27, 22, 11, 7, and the terminal claw about 5. The 
limb is rather liberally covered with very minute setae, which are arranged in rows. The 
lateral line is distinguishable throughout. The left femur of the first pair of legs and the 


right femur of the last pair bear peculiar excrescences, obviously the result of some injury. 
In both cases, however, the injured joint is longer than its fellow on the opposite side. 

A single specimen, which carries a few individuals of Scalpellum, was taken at the 
South Orkneys in June 1903. 

Colossendeis leptorhynchus. 
Colossendeis leptorhynchus, Hoek, (15), pp. 64-65. 

A specimen referable to this species was taken in lat. 48° 6'S., long. 10° 5' W., in 
1742 fathoms. It differs slightly from the type specimens taken by H.M.S. Challenger, 
and is much larger than any from that expedition. With its legs straightened out 
it covers an area of very nearly 1 4 inches. Its length is as follows : — Proboscis, 
39 "5 mm. ; trunk, 13 '5 mm. ; abdomen, 5 mm. ; total, 58 mm. 

The Body is perfectly smooth, but traces of segmentation may be seen under a good 
lens. It is not stoutly built, and the small lateral processes are separated by an interval 
equal to about half their thickness. 

The Cephalon is marked off from the rest of the trunk by a V-shaped groove which 
cuts into the space between the first pair of lateral processes. The ocular tubercle is 
immediately in front of this groove, and occupies rather a large area ; it is of very small 
elevation and bears two poorly developed eyes. 

The Proboscis is very long and slender, slightly enlarged in the middle, exactly 
as described by Dr Hoek ; but it is movably articulated to . the trunk as in the type 
and all other species of the genus that I have seen, twenty-three in number. 

The Palps arise ventro-laterally as close as possible to the proboscis, beyond which 
they do not extend far, the sixth joint not reaching the extremity. The first two joints 
are very small, the proportions of the remainder being 14, 2, 20, 2*5, 2 "3. The three 
terminals are together not as long as the preceding joint ; the first two are subequal, 
and the last a little longer, but not as long as the two together. The entire organ is 
very finely setose throughout. 

The Ovigers rise ventrally close together, each on a small body-process which is 
close behind the proboscis. They are very long ; the first three joints are as usual very 
small, the middle one of these being much more slender than the other two. The 
proportions of the three following are as 23, 4, 26. The four terminal joints are small 
and form the characteristic loop, and are provided with groups of denticulate spines, 
the last joint bearing a strong claw. A lateral line is readily seen along the entire 
appendage. There are not less than seven rows of these denticulate spines ; those of 
the first and principal row are the longest and best suited for observation. Here they 
are less numerous than in the other rows. They are roughly cylindrical shafts, becom- 
ing, at about half their length, flattened ovoid blades. The margin of the flattened 
blade is, near the base, provided with very small and rather curved teeth. These 
become longer and closer set, and before the extremity of the blade is reached they 


have become fine setae. I believe that this appearance is entirely due to wear and 
tear ; these spines and the terminal claw are more or less worn, some of course very 
much more than others. If in really good condition, I believe these spines would be 
finely dentate all round the blade. 

The Legs are long, very slender, and attain a length of 180 mm. The three coxae 
are small and subequal ; the proportions of the remaining joints are as 54, 61, 44, 8, 4. 
The terminal claw is very small. The limb is supplied at long intervals with very 
minute setae ; these make their appearance on the femur. The relative sizes of the 
joints of the leg differ from Dr Hoek's specimens, in which the femur is the longest 
joint. There can, however, be no doubt as to the identity of this species. 

In the preparation of this report I am greatly indebted to the Council of the Marine 
Biological Association of the United Kingdom, and especially to Dr E. J. Allen, the 
Director, for accommodation at their Plymouth Laboratory ; also to my friends Mr and 
Mrs L. E. Sexton and Mr J. Ritchie, who have so generously assisted me with 
the drawings and the photographs from which they have been prepared. 


(1) P>ohm, R., "Uber die Pycnogoniden des Konigl. Zoologischen Museums zu Berlin, insbesondere 
uber die von S.M.S. Gazelle mitgebrachten Arten," Monatsber. Akud. d. Wiss. zu Berlin, pp. 170-195, 
2 pi., 1879. 

(2) Bouvier, E. L., " Observations preliminaries sur les Pycnogonides recueillis dans la Region 
Antarctique par la Mission du Francais," Bull, du Mus, d'Hist. Nat., 1905, No. 5, pp. 294-7. 

(3) Bouvier, E. L., " Nouvelles observations sur les Pycnogonides recueillis dans les Regions 
Antarctiques au cours de la campagne dirig^e par M. Jean Charcot," Comptes rendus des Seances Acad, des 
Sciences, cxlii., 1906, pp. 15-22. 

(4) Buchan, A., "Report on Oceanic Circulation, based on the Observations made on board H.M.S. 
Challenger, and other Observations," H.M.S. " Challenger" Reports: Summary of Results, 1895. 

(5) Cole, L. J., " Pycnogonids of the West Coast of North America," Hurriman Alaska Expedition, 
vol. x., 1904, pp. 249-298, pi. ix.-xxvi. 

(6) Cole, L. J., "Tendegged Pycnogonids, with Remarks on the Classification of the Pycnogonida," 
Ann. and Mag. N.H., 7, vol. xv., 1905, pp. 405-415. 

(7) Dohrn, Anton, Die Pantopoden des Golfes von Neapel und der angrenzenden Meeresabschnitte, iii., 
Leipzig, 1881. 

(8) Eights, J., "Description of a New Animal belonging to the Arachnides of Latreille, discovered in 
the Sea along the Shores of the New South Shetland Islands," Jour. Boston Nat. Hist. Soc, vol. i., 1834-7, 
pp. 203-6. 

(9) Hodgson, T. V., "Crustacea," Report on the Collections of Natural History made in the Antarctic 
Regions during the Voyage of the " Southern Cross," 1902, pp. 257-8. 

(10) Hodgson, T. V., " On a new Pycnogonid from the South Polar Regions," Ann. and Mag. N. H., 
7, vol. xiv., 1904, pp. 458-462, 1 pi. 

(11) Hodgson, T. V. "Scotia Collections: On Decalopoda australis, Eights — an old Pycnogonid re- 
discovered," Proc. Roy. Phys. Soc. Edinburgh, vol. xvi., 1905, pp. 35-42, 2 pi. 

(12) Hodgson, T. V , "Decalopoda and Colossendeis," Zool. Anz., Bd. xxix., 1905, No. 8, pp. 254-6. 

(13) Hodgson, T. V., "Pycnogonida," Natural History Collections of the "Discovery" : Zoology, 
vol. iii., 1907. 

(14) Hodgson, T. V., " Pycnogoniden," Ergebn. Hamburg. Magalhaens Sammelreise, 1907. 



(15) Hoek, P. P. O, "Report on the Pycnogonida dredged by H.M.S. Challenger during the -years 
1873-76," H.M.S. " Ckallmqer" Reports: Zoology, vol. iii., 1881. 

(16) Hoek, P. P. O, "Nouvelles Etudes sur les Pycnogonides," Arch. Zool. Exp. et Gen., ix., 1881, 
pp. 437-542, pi. xxiii.-xxx. 

(17) Hoek, P. P. C, " The Pycnogonida dredged in the Faroe Channel during the Cruise of H.M.S. 
Triton (in August 1882)," Trans. Roy. Soc. Edin., vol. xxxii., 1883, pp. 1-10, 1 pi. 

(18) Hoek, 1'. P. C, "On Four Pycnogonids dredged during the Cruise of the Challenger," Tijdschr. d. 
Ned. Dierh. Vereen. (2), v., 2-4, 1898, pp. 290-300, 2 pi. 

(19) Loman, J. C. C. " Decolopoda, Eights, oder Colossendeis, Jarz," Zool. Anz., Bd. xxviii., 1905, 
pp. 722-3. 

(20) Miers, E. J., "Crustacea of Kerguelen Island (Pycnogonids, pp. 211-214)," Phil. Trans. Lond., 
vol. clxviii., 1879, pp. 200-214. 

(21) Mobius, K., " Arktische and subarktische Pantopoden," Fauna Artiea : Erne Zusammenstelhmg 
der arhtischen Tierformen, mil besonderer Besiichsichtigung des Spitzbergen Gebeites auf Grund der Ergebnisse 
der deutschen Expedition in das Nordliche Eismeer im Jahre 1898, Band ii., 1901, Lief. 1, pp. 35-64. 

(22) Mobius, K., "Die Pantopoden der deutschen Tiefsee Expedition, 189S-99," Wissensehaftliche 
Ergebnisse der deutschen Tiefsee Expedition auf dem Dampfer " Valdivia," 1898-99, Band iii., Lief. 6, 1902, 
pp. 175-196, Taf. xxiv.-xxx. 

(23) Pelseneer, P., "Mollusques," Resultats du Voyage du S.Y. " Belgica" en 1897-99, 1903. 

(24) Pfeffer, G., " Zur Fauna von Sud Georgien," Jahrb. d. Hamburgischen Wiss. Anst., Bd. vi., Heft 
ii., 1889, pp. 41-49. 

(25) Sars, G. O., "Pycnogonidea," Zoology of the Norwegian North Atlantic Expedition, 1891. 

(26) Schimkewitsch, W., " Sur les Pantopodes recueillis par M. le lieutenant G. Cherchia pendant le 
voyage de la corvette Vettor Pisani en 1882-85," Atti della R. Accad. dei Lincei, Mem., (4), vi., 1889, 
pp. 329-47, 1 pi. 

(27) Schimkewitsch, W., " Compte rendu sur les Pantopodes : Reports on Dredging Operations . . . . 
U.S. Fish. Comm. S.S. Albatross during 1891," Bull. Mus. Comp. Zool., vol. xxv., 1893, pp. 27-43, 2 pi. 

(28) Schimkewitsch, W., " Leber die Periodizitat in dem System der Pantopoda," Zool. Anz., Bd. xxx., 
1906, pp. 1-22. 

(29) Stebbing, T. R. R., "The Nobodies: A Seafaring Family " Knowledge, vol. xxv., 1902, chap, iv., 
pp. 185-89. 



Plate I. 



Ch xtonymphon assimile 

oviger x 15. 


3a. Nymphon lonyicoxa . 

$ palp x 10 



>' n 

9 palp x 16. 


4. Nymphon articulare . 

? palp x 50 



Nymphon capense 

palp x 20. 



oviger x 28. 



ii )> 

oviger x 20. 


5. Nymphon compactum 

oviger x 8. 



Nymphon longicoxa . 

oviger x 9. 



cJ palp x 1 5 

Plate II. 



Ainmothea communis . 

$ palp x 35. 


3. Colossendeis orcadense 

nat. size. 



n ;j 

oviger x 40. 


4. Pal/enopsis lanata 

oviger x 14. 



C/uxtonymphon orcadense 

$ palp x 13. 


4a „ ., 

c? palp x 16 



!) !! 

oviger x 12. 

Plate III. 


i . 

Leionymphon Clausi . 




2a. Delcalopoda australis 

proboscis from 



»» ?) 


oviger x 8. 

right side x 4. 


Decolopoda australis . 


nat. size. 



2c. „ „ 

oviger, terminal 

joints x 6. 
cheliferus x 6. 

o. i\uy. uuu.Liuni, 

Vol. XLVI. 

Hodgson: Pyc ono go nids — Plate 


M' FsrlBns b 6isk,i.cUH) Edin*" 

'"ns.Roy. Soc.Edm r 

Hodgson: Pyc on ogonids.— Plate II. 


M' fc BrskineXiBi Edirf 

ns.Roy. Soc.Edm r 


Hodgson: Pyconogonids— Plate III. 

M?Farl*n. kErskinaLitkEdiT 


( 189 ) 

VII. — Scottish Rotifers, collected by the Lake Survey (Supplement). 
By James Murray. (With Two Plates.) 

(MS. received November 8, 1907. Read November 18, 1907. Issued separately February 10, 1908.) 


In June 1906 there was published an account of the Rotifers collected in the 
Scottish lochs up to that date (7). Many Rotifers had been observed during the 
survey of the lakes which are not referred to in that paper, as they were not collected 
actually in lakes. It is with the object of recording these species, and especially of 
giving an account of some interesting forms previously unknown or imperfectly known,, 
that this supplementary paper is written. These Rotifers were, almost without excep- 
tion, collected from moss, and, as is always the case in these circumstances, the 
Bdelloida greatly preponderated over the other orders. 

Lists are given of the species found in the various islands visited — North Uist, 
Orkney, Shetland, — as an assistance to students of distribution, though there is little of 
special interest in these lists. The list from North Uist contains none but very 
common species, but in Orkney and Shetland there were a number of rare and interest- 
ing species. 

Two animals have been found with sufficient frequency on mountain-tops to suggest 
that they may be specially adapted to the climate of such situations. They are 
Callidina cornigera and C. plicata, var. hirundinella. Both extend into the most 
northerly lands — the former to Spitsbergen, the latter to both Spitsbergen and 
Franz Josef Land. 

Rotifers of the Scottish Lochs — Additional Species. 

Since the previous list was issued (7), nine additional species have been observed in 
our lakes, and one distinct variety : — 

Callidina tridens, Milne. Loch Tay ; St Mary's Loch. 

C. constricta, Dig. Frequent. 

C. musculosa, Milne. Loch Ness. 

Rotifer hapticus, Gosse. A stout brown animal with stout antenna, common in 
Loch Ness, seems to answer sufficiently to Gosse's description. The heavy antenna, 
narrowed towards the base, resembles that of Callidina armata, Murray (6). It is 
placed very far forward on the head in the creeping attitude. It is terminated by 
what appears to be a single seta, but may be a pencil of setae. Otherwise the animal 
has a strong resemblance to R. macroceros, Gosse, and I have a suspicion that it is that 


species with the antenna partly retracted. Gosse was evidently aware of the close 
affinity of the two species, as his description aims at discriminating them. 

Synchmta grandis, Zach. Castle Semple Loch, Kenfrew. 

S. oblonga, Ehr. Frequent in small lochs and reservoirs. 

Diaschiza era, Gosse. Loch Ness. 

Brachionus urceolaris, Ehr. Lochrutton, Dumfries. 

B. angularis, Gosse. Several small lochs near Dumfries. 

B. pala, Ehr., var. dorcas (Gosse). Hogganfield Loch, near Glasgow. 

Rotifers found in Ponds and Ditches. 

This short list is not a complete account of the Rotifers collected by the Lake Survey 
in ponds, but only of those pond species which did not also occur in lakes. A 
large number of species were got in ponds, including the majority of the species 
recorded for the Scottish lochs. Callidina natans, a pond species, has already been 
recorded (6). 

Floscularia cornuta, d'Ud. Bog pool on Blantyre Moor, near Glasgow. 

Philodina convergens, sp. n. (Plate I. figs. 1 to 3). 

Specific characters. — Of moderate size ; diameter of corona equal to or slightly 
exceeding first cervical segment ; collar prominent, dorsal folds ending below summit 
of upper lip ; sulcus between discs narrow, bottom convex, central setae on discs. Rami 
much constricted at teeth, with prominent spine behind ; teeth 2/2. Antenna short. 
Foot four-jointed. Spurs moderately long, with straight interstice, expanded at base ; 
upper part almost parallel-sided, abruptly acute. Dorsal toes much smaller than 

Length, feeding, 240 to 280 /x ; creeping, 333 m. Colour reddish. Trunk closely 
longitudinally plicate. Diameter of corona, 50 m ; jaw, 25 m long ; spurs, 22 to 25 m 
long. The rostrum is short, and no long tactile setae were seen. Length of antenna 
equal to half the diameter of the first cervical segment. 

Pale brown eyes. Stomach walls with yellow globules. 

Habitat. — In washing of moss from the river Lochy, Killin, January 1906; 
quarry at Nerston, East Kilbride, February 1906. 

In both localities the washing contained many Gammarus, and Gammarus 
parasites (at Killin Phil, hamata, at Nerston P. laticeps), so there is at least a 
suggestion that this is also a parasite. The corona is usually carried with the discs 
approximated as in P. laticornis, but can be pretty widely extended. 

The points which distinguish this from other Philodime may seem rather slight, 
but close attention to them will show that it is quite distinct. The spurs, though 
narrow, are of the same form as in P. laticeps, etc., that is, blade-shaped. All the 


species of the central group have more gradually tapering spurs. It is nearest P. 
laticornis, from which it is distinguished by the much shorter foot, slender spurs, and 
oviparous reproduction. The marked constriction of the jaws above the middle, and 
the posterior spine, are shared by no other species. All have a prominence at the back 
of the ramus, and in some it might be called a tooth or blunt spine, but none have 
such a prominent spine. This is only exceeded in size, among species known to me, in 
an undescribed African Callidina. 

Rotifer macrurus, Schrank. Common in the mud of ponds, and especially so m 
peaty bog pools. 

Proales werneckei, Ehr. Parasite in Vaucheria, on which it forms galls ; roadside 
ditch, Blantyre Moor, near Glasgow. 

Diglena rosa, Gosse. Ponds, Pomona, Orkney. 

Rattulus ca,rinatus, Ehr. Peaty pools, Blantyre Moor. 

R. bicristatus, Gosse. Pools, Blantyre Moor. 

Elosa worrallii, Lord. Pools, Blantyre Moor. 

Stephanops muticus, Ehr. Pools, Blantyre Moor. 

Diaschiza exigua, Gosse. Pond near Mallaig, Inverness-shire. 

Pterodina mucronata, Gosse. Pond at Nerston, near Glasgow. 

Rotifers found among Moss. 

Moss in any situation is likely to contain abundance of Rotifers. In most situations 
Bdelloids will predominate, but Sphagnum will harbour many other kinds. As a rule 
the Bdelloids are indifferent to the situation of the moss, and many kinds are found 
alike in moss which is always moist [Sphagnum and mosses in streams, etc.), and in 
that which is only intermittently moist (on the ground, walls, trees, rocks, etc.). 

A few species are exceptions to this rule. Microdina paradoxa, for example, will 
only be found in fresh or running water. The symbiotic species, supposed to be 
confined to certain hepatics, are quite common in other situations, as among Sphagnum. 

Callidina microcephala, Murray. Ground moss, Fort- Augustus. 

Philodina brycei (Weber) (8). 

One of the commonest Bdelloids in Scotland is P. brycei, and it is often found in 
lochs. Tt is subject to a great deal of variation. 

Weber's description of Callidina brycei (8, p. 347) is in the following terms : — " Le 
corps est robuste, allonge, de coloration grisatre ou brunatre. La peau est rugueuse, 
sillonnee de forts plis longitudinaux sur les faces dorsale et laterales du corps ; par 
contre, on compte 8 plis transversaux sur la face ventrale. Le bord anterieur du 
premier segment troncal et le bord posterieur du troisieme segment troncal sont armes 
chacun, dorsalement et lateralement, d'une rangee de courtes epines. L'organe rotatoire 
est moderement large, a sillon intertrochal etroit. La trompe cylindrique est longue. 


L'antenne dorsale est longue et formed de deux articles. Le pied est court, epais. Les 
eperons courts, coniques sont peu distants l'un de l'autre. Formule dentaire 2/2." 

Wkber remarks tlie resemblance to Callidina alpiinn, Ehr. (which I transfer to 
Philodina). The resemblance is extremely close. Size, colour, general form, skin-folds, 
— all are alike in the two species. Imagine P. brycei deprived of its short central and 
anterior spines, and we could not, I believe, distinguish it from P. alpium. 

The type of the species has never occurred in our collections in Scotland, but several 
varieties are common. The common Scotch form which comes nearest the type differs 
in having two short spines on the posterior part of the trunk. These appear to be on 
the preanal segment, but I believe they are really on the fourth central. 

The spines on the anterior border of the trunk do not form a regular equal series, as 
in Weber's figure. The pair nearest the middle of the back are longest, and form a 
fork in which the antenna rests when the animal is feeding, precisely as in Brachionus, 
Anursea, etc. 

The next pair of spines are very small points. The last pair are laterally placed, or 
a little inclined to the ventral side. Each of them is usually furcate, one point (the 
lateral one) being erect, and the other (more ventral) spreading or decurved. 

There are some other small spines usually present, which are not referred to by 
Weber. On the two lateral skin-folds of each side of the central trunk, on which are 
the last spines of the transverse dorsal row, there are small spines a little in front of 
those. Sometimes only one lateral fold has this spine. 

This form is generally distributed in Scotland. It is not figured here, but fig. 13, 
which is of another variety, indicates all the structures referred to, but has an 
additional transverse row of spines. 

Variety (fig. 13). — This differs from the form just described in possessing a second 
transverse row of spines on the central trunk, a little behind the main row. There 
are four spines in this row. The anterior processes forming the fork for the antenna 
are very large and are frequently furcate. Almost as common in Scotland as the 
other, and widely distributed over the world. In Indian examples there may be six or 
eight spines in the secondary transverse row. 

Callidina cornigera, Bryce (2). 

In 1893 Mr Bryce described this species in the following terms (p. 201): — 
' Trochal discs apparently without gap, laterally produced into two horn-like but 
fleshy processes, whose bases are furnished on inner face with cilia, forming part of 
principle wreath. Antenna very short, one-fourth of neck thickness. . . . Length, 
extended, about TO^th inch." He noticed further that "the double flap terminating 
the column tip (lamellae) was rather more developed than usual," saw the usual skin- 
folds and conical spines, but failed to make out the dental formula. 

These observations were made on a single example, which, though kept alive for 


the long period of fourteen days, appears never to have given very good opportunity 
for study. 

The original specimen was found among moss from roadside, near Bognor, Sussex. 
Some years later another example was found in moss from Buckinghamshire. In 1897 
a third example appeared in moss from Spitsbergen. On this occasion Mr Bryce 
observed the dental formula 2/2, and gives the measurement of the extended animal as 
0347 mm., but adds nothing further to his original description. In the plate 
accompanying the description the jaws are figured with two teeth in each. 

Weber (8) found two examples of the species, in very bad condition. His descrip- 
tion (p. 349) is almost identical with Bryce's, and he adds nothing of importance 
except the dental formula 2/2. He says that the lamellae are broad. 

I can find no other record of the species, which would seem to indicate that it is 
a rare animal, as such an extraordinary creature would readily attract attention. In 
March 1904 I first found it in moss from the shore of Loch Ness, near Fort- Augustus, 
and subsequently it has appeared pretty frequently in Scotland, permitting fuller 
studies to be made of it. 

There will hardly, I think, be any doubt as to the identity of the animal found in 
Scotland, if the figures here given (Plate II. figs. 20 to 26) are compared with those of 
Bryce and Webeu, although I am compelled to differ from Mr Bryce as to the most 
important structures of the animal, the horns and the discs. According to Bryce, 
the discs have apparently no gap, and are produced into the horns, and the bases of 
the horns are ciliated, the cilia forming part of the principal wreath. 

I find that the discs are normal and separate, but close together, and the horns 
spring from folds of skin somewhat to the dorsal side of the discs (technically part of 
the collar, I believe). The horns are nowhere ciliate. 

These differences between Mr Bryce's observations and mine seem greater on paper 
than they are in reality. Compare the figures, and see what a slight difference in 
interpretation is involved. 

A full description of the animal, as observed in Scotland, is added. 

General form. — Slender, more like Weber's figure than Bryce's, broadest in central 
trunk, well-marked neck, slightly expanded head, much contracted in posterior trunk 
(4th central), foot generally hidden. Trunk closely longitudinally plicate. In its 
movements the animal contorts its body a good deal, changing form greatly, expanding 
one segment and drawing in another. 

Discs. — Elliptical, touching at inner margin, slightly inclined forward. 

Horns. — Long, white, with broad bases and narrow, soft, blunt tips, curved forward 
from base to apex, and elbowed where the narrow part begins, sometimes angled again, 
or incurved close to the apex. One horn is generally longer than the other, and in the 
creeping attitude the tip of the longer one usually protrudes, sometimes both. 

There is no doubt that the function of the horns is tactile. The animal is exceed- 
ingly sensitive and timid. The horns are employed before beginning to feed, and 


their action is very like that of fingers. Arrived before some flocculent material, the 
head is unfolded in a hesitating manner, the horns (or fingers) make a tapping motion 
forward among the food-material, after which the beast either withdraws into its trunk 
hurriedly, or goes on quietly feeding. The fingers seem to be moved by special 
muscles, and act as if articulated at the bend, though nothing of this was actually seen. 

Rostrum. — The basal joint is large, the apical joint very small and short. 

Lamellae. — These exhibit the highest development of the Bdelloid lamellae known. 
They are very large, and widely divergent laterally, resembling a pair of butterfly's 
wings. Unlike those of most Bdelloids, which remain passively motionless when 
extended, the lamellae of C. cornigera are waved about in the water with a very 
graceful motion. This action has suggested that they are organs of smell. 

Rostral cilia. — The brush of cilia is well developed, but none of the other processes 
of the tip were seen. 

Antenna. — This is very short. In the creeping attitude it looks like a little button 
(fig. 22), or may be quite retracted. It can be extended to about one-quarter the 
diameter of the neck. 

Stomach. — Walls much convoluted, food moulded into pellets. Though neither 
Bryce nor Weber mentions this, their figures seem to indicate that they observed it. 

Intestine. — Rarely easy to observe in a pellet-maker, it was here conspicuous and 
filled with pellets. 

Jaws. — Teeth 2/2. Lateral margin lightly convex, angled to anterior margin ; 
posterior margin abruptly bent, forming a little point (fig. 25). 

Foot. — Joints, four. Spurs small, acuminate, curved, close at bases. Toes, three, 
short, blunt. 

Yelk-mass. — Eight relatively large nuclei. 

Egg. — Small, thin-shelled, shortly oval. 

Habitat. — Among moss, most commonly Sphagnum, from near sea-level to nearly 
4000 feet. Shores of Loch Ness, several places ; Ben Lawers ; East Kilbride, near 
Glasgow ; Orkney, summit of Ward Hill, in Hoy, about 1500 feet. 

Length, creeping, 250 m. This is the same size as Mr Bryce's first example, but 
that from Spitsbergen was much larger. 

Callidina minuta, sp. n. (Plate I. figs. 11, 12). 

Specific characters. — Very small. Very short, with broad trunk. Corona less than 
neck, much less than collar. Central trunk broadly elliptical, first segment anterior to 
central trunk bearing some small spines on lateral skin-folds. Posterior part of trunk 
nodose. Foot short; spurs short, acuminate, meeting at bases. Teeth about 5/4 (5/5, 
4/3). Food moulded into pellets. 

In form this curious little animal is an exact miniature of C. pulchra, Murray (6). 
The only differences in detail which can be pointed out are that the first cervical does 


not project so prominently outward in C. minuta, and the little lateral spines on the 
anterior trunk segments. The skin was not stippled, but little importance can be given 
to that character. 

In size it does, however, differ greatly, measuring only 77 ^ in length when feeding, 
or little more than one-third that of C. pulchra, While size alone cannot be made a 
primary specific character, in this case the difference between the related species is so 
great that it gives greater weight to the other small differences. 

No Bdelloid which I have seen hatched is to such a degree smaller than the adult. 
In common with many other Rotifers, the young, when hatched, is hardly inferior in 
length to the parent, and very soon attains to all the adult proportions. The pellets 
of food which filled both stomach and intestine showed that the last example of C. 
minuta had been feeding for a considerable time. The pellets are relatively large, 
measuring 6-7 m- 

The discs are inclined backward, and have central setse. The spurs measure 9 n 
from tip to tip. 

The length of the antenna is equal to one- third of the diameter of the neck. 

Habitat. — Among Sphagnum, Blantyre Moor, 1902. In ground moss, Nerston, 
near Glasgow, March 1905. 

On the first occasion of finding it, my acquaintance with Bdelloids was very 
limited, and after a time I lost faith in the observation, and came to suppose that there 
had been a mistake as to the power of the microscope used. The second occurrence of 
the animal, years later, after a good deal of experience among Bdelloids, confirmed the 
earlier record. 

Callidina circinata, sp. n. (Plate I. figs. 4 to 10). 

Specific characters. — Small. Head nearly square, with corona slightly exceeding the 
collar, and very prominent dewlap overhanging first neck-segment. Upper lip of very 
unusual form — sulcus between discs deep, bounded by two large processes connected 
with the collar (the arrangement will be better understood from fig. 5 than from any 
description). Antenna short, dorso-lateral processes large, widely spreading laterally. 
Foot short, three-jointed. Spurs large, long, nearly parallel-sided, incurved, acute, inter- 
space large, flat. Toes symbiotic (joined to form a perforate disc). Teeth, 3/3 or 
2 + 1/1 + 2. Length vdien feeding, 213 m ; diameter of trunk, 71 m ; of corona 60 /* ; 
tip to tip of spurs, 30 m. 

Habitat. — Among moss growing on dry wall at Nerston, East Kilbride, October 
1906. Plentiful. 

This species affords another of those puzzles in distribution which are especially 
familiar to students of the Rotifera. The moss on the wall was growing in little 
detached cushions. One of these was casually pulled in passing, and was washed on 
reaching home. When the strange species was found so abundant in it, the wall was 
visited again, and surrounding tufts pulled, but none of the Rotifers found. 



The long spurs, more or less curved towards one another, resemble callipers. The 
flame-cells are narrow and spindle-shaped. The rostrum is short and broad. The 
dorso-lateral processes of the neck are conspicuous in the creeping attitude (fig. 6). 
Length of antenna about equal to half diameter of neck. 

The possession of a " symbiotic " foot, i.e. one ending in a perforate adhesive disc, 
distinguishes C. circinata from the great majority of Callidinse. Of the small group 
having a foot of this type, all have a greater number of teeth, except only C. symbiotica 
and C. armata, and both of these species have short spurs of quite a different shape, so 
that the species cannot be confused with any known species. 

Callidina plicata, Bryce ( 1 ). 

Described by Bryce in 1892. The most important distinguishing character is the 
expanded, hood-like, posterior trunk-segment. 

One of the very commonest Bdelloids in Scotland, C. plicata is the centre of a large 
series of forms, some of which have the processes of the posterior trunk-segment greatly 
produced, while others have them reduced or entirely absent. The latter would not be 
recognised as C. plicata, but long study of other characters of the type, and of a series 
of forms in which the processes were progressively reduced, convinced me that those 
without the processes were also of this species. 

Description of C. plicata (type) (fig. 14). 

General form narrow, elongate, widest in central trunk. Neck and anterior trunk 
well extended in usual feeding attitude. Trunk closely plicate, the central dorsal pair 
running out on to the rump, as pointed out by Bryce ; but the rump is marked off from 
the central trunk by a very deep transverse furrow, so that the longitudinal folds on 
the rump might more properly be regarded as distinct from those of the central trunk. 

The transverse furrow is much further forward than the lateral constriction separating 
rump and central trunk, but I believe both belong to the same line of separation of segments. 

Head. — From the neck the head gradually expands upward, the corona being the 
widest part. The collar is distinct, but not very prominent. The discs bear central 
setae, springing from small papillae. 

U pper lip. — Though to some extent variable, the upper lip is one of the best 
characteristics of the species. There are two convex processes, meeting in the middle 
line, much as in C. habita (3). Between them is usually a smaller convexity, not 
belonging to the lip. The processes are connected with the collar by a ridge, as usual. 
From the outer edge of the processes two fine lines pass downward towards the rostrum, 
converging towards the middle line, giving the lip the appearance of a keystone set into 
the head. This form of lip I have seen in no other species. 

Rostrum. — Short and broad, and bearing two lamellae, which in dorsal view seem 
far separated, as in some symbiotic Callidinse. The restlessness of the animal when not 


feeding has prevented a fuller study of the rostral processes ; hut I have thought at 
times I could detect long setae, or pencils of setae, projecting laterally from the lamellae, 
as in C. ehrenhergi, Janson (4). 

Vibratile tags. — These are of the usual narrow form. I have not been able to 
count more than four pairs. 

Alimentary canal. — The jaws, stomach, and intestine are normal. There are two 
teeth in each jaw, and the usual fine striae. The food is not moulded into pellets. 

Rump. — The preanal and anal segments, which form the rump, are but slightly 
separated ; but the second constriction mentioned by Bryce doubtless marks the 
separation, and the hood with its processes will then belong to the anal segment. The 
two lobes which project from the end of the anal segment are thin hyaline flanges. 

Foot of four joints, short. Spurs rather long for a Calliclina of the central group, 
narrow, tapering, slightly acuminate, obtuse, close together at base, but small interspace 
showing in some positions. Toes very short, difficult to see, as the last portion of the 
step is very rapidly made. 

Variations of C. plicata. 

Head. — The processes of the upper lip, which usually meet in the middle line 
(fig. 14), are sometimes separated by a small but deep sulcus. A ligule is very 
commonly present, in examples otherwise typical (see dotted line between the discs in 
fig. 15). In its greatest development this is a drop-shaped body, apparently just 
touching the head, a little to the ventral side of pedicels. 

Foot. — The first foot-joint sometimes has a boss, which may be centrally placed, as 
in most species having this process, or at the posterior edge of the segment, as in 
C. symbiotica. In an Indian form, not otherwise peculiar, this becomes a transverse 
ridge, with a sharp edge. 

Rump. — Bryce does not refer to the processes on the posterior margin of the hood, 
but his figure shows them. 

Their commonest form is shown (in fig. 14). A form in which they are quite 
obsolete is figured (fig. 19). The posterior part of the hood is divided into three plane 
surfaces, a central one and two lateral flaps. 

The processes are, on the other hand, often produced much more than in that 
figured (fig. 14), and become in some forms very prominent. They are still, however, 
quite rigid outgrowths of the skin, and do not appear to be articulated at the base. 

A form in which the development of these processes is carried still further is 
regarded as being a stable variety, and is described below. 

C. plicata, Bryce, var. hirundinella, var. nov. (Plate II. figs. 16 to 18). 

Distinctive characters. — The processes on the anal segment are produced into long 
ligular bands with rounded ends. They normally diverge widely, but they are 
distinctly articulated at the base, and can be approximated, or even crossed over one 



another (fig. 18). These tail-like processes suggest a resemblance to those of the 
swallow-tail butterfly. They are sometimes considerably longer than the foot. I have 
seen no evidence that they are movable at will, but they certainly yield readily at 
the base and assume many different positions. The position in which they are crossed 
does not seem one likely to be produced automatically by the movements of the skin 
to which they are attached. 

Habitat. — Among ground moss. Frequent, especially on mountain-tops. Fort- 
Augustus ; Ben Lawers (cairn on summit) ; North Uist ; summit of Ward Hill in 
Orkney, and of Ronas Hill in Shetland. 

Rotifer qvadrioculatus, Murray. Moorland, Fort-Augustus. 

Arthroglena lutkeni, Berg. Moorland, Pomona, Orkney. 

Stephanops microdactylus, Murray. Blantyre Moor ; Ballagioch Hill, near Glasgow. 

Cathy pna ungulata, Gosse. Mainland of Shetland. 

Brachionus bakeri, Ehr. Summit of Ronas Hill, Shetland. 

Rotifers found in the Sea. 

Though the Rotifers are pre-eminently a fresh-water group of animals, a considerable 
number of marine species are now known. Besides those which are exclusively marine, 
many species live indifferently in fresh or salt water. Seaweed was gathered on the 
west coast of Scotland, kept tightly packed in a tin box for a week, then washed in 
fresh water. A great many living Rotifers were then found which seemed active and 
healthy in the fresh water. 

The Lake Survey did not make a special study of the marine species, but occasionally, 
when opportunity offered, the nets were used in the sea, or seaweed was washed. 

Synchseta is a specially marine genus, or at least has a number of exclusively 
marine species. In the harbour at Mallaig we found one species extremely abundant, 
and in . the brackish loch of Stenness there were several species ; but we could not 
venture to name these without submitting them to a specialist. We have found only 
one Rotifer in the sea which has never occurred in fresh water. 

Rattulus dubius (Lauterborn) (5). In Morar Bay, West Inverness. Got by 
washing seaweeds. 

Furcularia reinhardti, Ehr. This common lake species was also frequent in the sea 
at Morar. The salt-water form was larger than the other, but I could see no other 

Philodina. Two species occurred in the washings of seaweed from Lochinver, West 
Sutherland. They could not be identified. (P. Jlaviceps, Bryce, was got by Mr Wm. 
Evans in the sea near Edinburgh.) 

Colurus, sp. In the washings of seaweed, one of the species occurring most 
regularly is a large Colurus, which I will not attempt to name, in the present chaotic 
state of the genus. 

Pterodina, sp. Lochinver. 



Rotifers of North Uist. 

The Rotifers enumerated in the following short list were collected in the early 
summer of 1904, while the lochs of the island were being surveyed. As the whole 
district surveyed is one great wet moor, most of the Rotifers were got from bog or moor 
mosses, chiefly Sphagytum and Hypnum of various species. 

One interesting collection was made from a tuft of Fontinalis growing in a mill- 
stream, within reach of the tides, so that the animals living in the axils of the leaves 
must be able to endure alternately fresh and salt water. The species found here were 
Philodina fiaviceps, and all four species of Euchlanis on the list. P. Jlaviceps has 
since been found actually in the sea, by Mr Wm. Evans. 

Philodina rugosa, Bryce. 

,, nemoralis, Bryce. 

,, jlaviceps, Bryce. 

„ macrostyla, Ehr., and variety tuberculata 
Callidina aspera, Bryce. 

„ lata, Bryce. 

,, angusticollis, Murray. 

„ sijmbiotica, Zel. 

„ tetraodon, Ehr. 

„ russeola, Zel. 

„ plicata, Bryce, and variety Idrundinella. 

„ papillosa, Thompson. 

,, multispinosa, Thompson. 

,, quadricornifera, Milne. 
Rotifer tardus, Ehr. 

,, macrurus, Schrank. 
,, citrinus, Ehr. 
Adineta vaga, Davis. 

Adineta barbata, Janson. 
Ascomorpha ecaudis, Perty. 
Microcodon claims, Ehr. 
Copeus cerberus, Gosse. 
Proales petromyzon, Ehr. 
Furcularia longiseta, Ehr. 
Rattulus longiseta, Schrank. 
Diurella tigris, Miill. 
Diaschiza gibba, Ehr. 
Dinocharis tetractis, Ehr. 
Stephanops muticus, Ehr. 
Euchlanis lyra, Huds. 

,, oropha, Gosse. 

„ deflexa, Gosse. 

,, triquetra, Ehr. 
Monostyla lunaris, Ehr. 

„ cornuta, Ehr. 
Colurus obtusus, Gosse. 
Anurxa serrulata, Ehr. 

Rotifer a of Orkney and Shetland. 

As I am aware of no published account of the Rotifers of these islands, a list of some 
of those found by the Lake Survey is here given, though there is nothing of special 
interest in the list. We found several species which are very local in their distribution, 
some of them only previously known from one or two spots on the mainland of 
Scotland. Though the majority of Rotifers are considered to be cosmopolitan in their 
distribution, the same species appearing under similar conditions everywhere, the portion 
of the earth's surface which has been carefully examined is too limited to permit us to 
suppose that this is demonstrated, and so local lists have their uses. 

Arthroglena lutkeni, Berg., occurred in Orkney, and Brachionus bakeri, Ehr., in 
Shetland, both for the first time in my experience. The lists from Orkney and Shetland 
are given in parallel columns, though the information thus given is of little value, owing 
to the inadequacy of the work done. As usual, the great majority of the species found 
are Bdelloids. Forty-four species and three varieties are noted for Shetland, thirty-two 



species and one variety for Orkney. Twenty-three species and one variety are common 
to the two groups. 

The two highest hills in the islands, Ward Hill in Hoy and Ronas Hill in Shetland, 
were specially examined. Both are about 1500 feet in height, and thus combine high 
latitude and considerable elevation. Saxavord Hill, in Unst, nearly 1000 feet high, and 
close to the northernmost point of the British Islands, was also carefully studied. 



Philodina aculeata, Ehr. .... 

Ronas Hill, Mainland. 

,, macrostyla, Ehr. 


Mainland; Yell. 

var. tvberculata (Gosse) . 


Mainland; Yell. 

,, acutieomis, Murray . 

Ronas Hill ; Yell. 

,, brevipes, Murray 

Ward Hill," Hoy! 

Ronas Hill. 

„ flaviceps, Bryce 


,, nemoralis, Bryce 

Hoy ; Rousay 


,, rugosa, Bryce .... 

Ward Hill, Hoy. 

Ronas Hill ; Yell. 

„ alpium (Ehr.) .... 

Saxavord, Unst. 

,, brycei (Weber) .... 

Rousay .... 


„ humerosa, Murray 


Callidina aspera, Bryce .... 

Rousay .... 

Mainland ; Yell ; Unst. 

,, roeperi (Milne) .... 


,, angusticollis, Murray . 

Saxavord, Unst. 

,, lata, Bryce .... 

Ronas Hill. 

,, pulchra, Murray 


,, const ricta, Duj. .... 

Ward Hill, Hoy. 

,, tridens (Milne) .... 


,, leitgebii, Zel. .... 

Hoy ; Rousay 

Saxavord, Unst. 

,, cornigera, Bryce 

Ward Hill, Hoy. 

,, plicata, Bryce .... 

Hoy ; Rousay . 

Mainland ; Yell ; Unst. 

var. hirundinella, var. no v. 

Ward Hill, Hoy . 

Ronas Hill. 

„ habit a, Bryce .... 

Ward Hill, Hoy 


,, quadricomifera (Milne) 

Pomona ; Rousay 

Mainland; Yell; Unst. 

,, ehrenbergii, Janson . 

Pomona .... 

Yell; Unst. 

,, 2 m piU° sa > Thompson . 

Hoy ; Pomona ; Rousay 


,, muscidosa (Milne) 

Pomona .... 


,, symbiotica, Zel. 

Pomona .... 

Ronas Hill. 

,, armata, Murray 


„ tetraodon, Ehr. .... 

Pomona .... 

Ronas Hill. 

,, incrassata, Murray 

Rousay .... 


,, scarlatina, Ehr. 

Ward Hill, Hoy . . 

Ronas Hill. 

„ russeola, Zel. .... 

Ward Hill, Hoy 

Mainland; Unst. 

Rotifer vulgaris, Schrank .... 

Yell; Unst. 

,, macrurus, Schrank 

Ronas Hill. 

,, citrinus, Ehr. .... 

Saxavord, Unst. 

„ longirostris (Janson) 

Ward Hill, Hoy '. 

Mainland ; Unst. 

Adineta vaga (Davis) .... 

Pomona .... 


var. major, Bryce 



,, barbata, Janson .... 

Ward Hill, Hoy 

Mainland ; Unst. 

,, gracilis, Janson .... 

Pomona .... 


„ tuberculosa, Janson 

Ronas Hill. 

Notommata toruloso (Duj.) . 


Diglena rosa, Gosse . . • . . 



,, ferox, Western .... 

• *.... 


Arthroglena lutkeni, lierg. 



StepJimiops stylatus, Milne 


„ tenellus, Bryce 

Pomona .... 




Diaschiza eva, Gosse 
Monostyla lunaris, Ehr. 

„ cornuta, Ehr. . 
Distyla flexilis, Gosse 
Cathypna ungulata, Gosse 
Colurus obtusus, Gosse 
Metopidia acuminata, Ehr. 
Brachionus bakeri, Ehr. . 


Ward Hill, 



Ronas Hill. 

Ronas Hill. 
Ronas Hill. 


(1) Bryce, D., "On the Macrotrachelous Callidinse," Journ. Quekett Micr. Club, Ser. II., vol. v., 

1892, p. 15. 

(2) ,, "On Two New Species of Macrotrachelous Callidinse," Journ. Quekett Micr. Club, 

Ser. II., vol. v., 1893, p. 196. 

(3) ,, " Farther Notes on Macrotrachelous Callidinse," Journ. Quekett Micr. Club, Ser. II., 

vol. v., 1894, p. 436. 

(4) Janson, 0., Die Rotatorien-Familie der Philodineen, Marburg, 1893. 

(5) Lauterborn, R., "Die pelagischen Protozoen u. Rotatorien Helgolands," Wiss. Meeresuntersucli. 

Biol. Anstalt auf Helgoland, 1894, p. 207. 

(6) Murray, J., "A New Family and Twelve New Species of Rotifera," Trans. Hoy. Sor. Edin., xli., 

1905, p. 367. 

(7) „ "The Rotifera of the Scottish Lochs," Trans. Roy. Soo. Edin., xlv., 1906, p. 151. 

(8) Weber, E. F., "Faune Rotatorienne du Bassin du Leman," Rev. Suisse de Zool., v., 1898, p. 264. 


In order to indicate relative sizes of different species, all the drawings of complete animals are made 
to one scale, except that of C. minuta. To show sufficient detail, this had to be drawn nearly twice as large 
as the others. 

Plate I. 

1. Philodina convergens, sp.n., dorsal view, feeding. 

2. ,, ,, jaw, seen from behind. 

3. ,, ,, spurs and toes. 

4. Callidina circinata, sp. n., dorsal view, feeding. 

5. „ ,, head, morehighlymagnified. 

6. „ ,, head and neck, when creep- 


7. Callidina circinata, section of neck and antenna. 

8. „ „ jaw. 

9. ,, „ foot. 

10. ,, ,, spurs and perforate disc. 

11. Callidina minuta, sp. n., dorsal view, feeding. 

12. ., „ head more highly magnified. 

13. Philodina brycei, Weber, variety. 

Plate II. 

14. Callidina plicata, Bryce, type, feeding. 

20. Callidina cornigera, Bryce, side view, feeding. 

15. ., ,, head, on larger scale. 21. 

16. ,, ,, var. hirundinella, var. nov., 22. 

rump and foot, showing 23. 


17. ,, ,, var. hirundinella, side view 24. 

of foot. 

18. „ „ var. hirundinella, showing 25. 

processes crossed. 26. 

19. ,, ,, foot of variety without 

flanges on hood. 


dorsal view, feeding, 
head, seen from above, 
side view of rostrum, 

lamellae, and antenna, 
dorsal view of rostrum and 




12 M. 1208 

Trans. Roy. Soc. Edm r 


Scottish Rotifers. 

Plate I. 


lirrray, i.l. ^ „ 

),P H ,LOD 1N A CONVERGED. sp .„ 4"10, CALLID1NA C.RCiNATA, .,.„. lH2,C.NlKUTA,.p.,, 13,P BRYCEL WEBER, 

iS.Roy. Soc.Edm r 

Murray: Scottish Rotifers. Plate II 


M' Farias 8, ErskinelitK .Elfin* 

419.CALLIDINA PLICATA,Bry Ce . 16 18 

r & v. HIRUNDINELLA.var.nov. 20-26,C.C0RNIGERA,Br 

( 203 ) 

VIII. — On the Histology of the Ephedreae, with Special Reference to the Value of 
Histology for Systematic Purposes. By R. J. D. Graham, M.A., B.Sc, Carnegie 
Research Scholar, Botanical Department, The University, St Andrews. Com- 
municated by R. A. Robertson, M.A., B.Sc, F.R.S.E., F.L.S. (With Three 

(MS. received April 30, 1907. Eead June 10, 1907. Issued separately February 20, 1908.) 

Part I. — Primary Structure of Stem. 

Although a few papers have been published a number of years ago on the genus 
Ephedra, the authors have treated the subject mainly from a systematic standpoint. 
Thus, while the characters of the floral organs have been carefully examined and 
described, the structure and character of the vegetative organs have only received 
passing reference. It is proposed in the following notes to undertake the study of the 
histology of the vegetative organs, and thereafter to ascertain how far this knowledge 
applies to the determination of species. The following notes deal with the histology 
of the primary stem. 

The material used, including in all 16 species and varieties, was kindly supplied 
from the following sources : — 11 species from the Director of the Royal Botanic Garden, 
Kew ; 8 species from the Director of the Royal Botanic Garden, Edinburgh ; 2 species 
from Professor Trabut, Algiers ; E. Helvetica from Professor Schroeter, Zurich ; 
E. altissima Helvetica from Professor Flahault, Montpellier, — to all of whom I here 
tender my thanks. Material for the study of 3 species was purchased from the 
Botanical Supply Association. For permission to examine the living specimens in the 
Royal Botanic Gardens of Kew and of Edinburgh I have to thank the Directors of these 

(Ecology. — The genus Ephedra, consisting of some 32 species, belongs to the 
Gnetacese, an order of the Gymnospermse. The majority of the species are found in 
the desert parts of both the Old and New Worlds. Correlated with their physiologically 
dry surroundings, the plants exhibit to a very marked degree various xerophilous 
adaptations. The leaves in almost every species are reduced to mere scales, while the 
green colour of the stem shows that the photo-synthetic functions are carried on there. 
Further reduction of transpiration is effected through the sinking of the stomata in 
furrows, while each stoma lies in a deep depression of the epidermis. The tender apex 
is surrounded by many envelopes of leaves, being thus protected from excessive 
insolation. At the approach of the dry season many species lose a great number of the 
young branches, thereby leaving only the more mature stems to function through the 
dangerous period of the year. The erect broom-like habit, so characteristic of many 
species, exposes a minimum area to the strong light of the noonday sun, while the 


204 MR R. J. D. GRAHAM ON 

position enables full advantage to be taken of the optimum light of the early and the 
late day. 

Tegumentary System. — The surface of the stem of the Ephedreae closely resembles 
that of an ordinary Equisetum, in possessing longitudinal ridges and furrows (Plate I., 
fig. 15), the former corresponding in position with hypodermal stereom strands. 
The cells of the epidermis are rectangular or slightly hexagonal in shape, being 
elongated in the direction of the stem axis. The cells at the base of the internode are 
not usually so much elongated as those in the middle portion of the same. The outer 
cell walls consist of three strata (Plate I., fig. l) ; the most external is a heavily cutinised 
stratum, bounding a middle layer of mucilage, containing small crystals or granules of 
calcium oxalate, which tend to have a stratified arrangement. This layer is limited 
internally by a thick stratum of cellulose. The cells contain the normal cell con- 
stituents, with, in some cases, starch grains, while again the cell contents may be 
stained various shades of yellow. The coloration is apparently due to the presence of 
tannin, as the cells give a black reaction on treatment with ferric acetate. In some 
cases the epidermal cells are divided into two by a wall which may be parallel, or nearly 
so, to any of the three dimensions of the cell. Either of the daughter-cells may further 
divide by a wall inclined at an angle to the first division wall (Plate I., fig. 2). In 
some species this division occurs more frequently in the neighbourhood of a stoma, 
though it may also be shown in independent cells. No trace could be found of the 
pores recorded by Stapf as passing through the cell wall. The only suggestion of their 
appearance was the striation in the cellulose layer seen after treating the section with 
sulphuric acid and methylene blue. These striae ended at the limit of the cellulose layer. 

The epidermal cells covering the ridges are, as a rule, larger than elsewhere ; while 
the external walls are either markedly convex outward, or bear on their outside papillae 
into the base of which a blunt protoplasmic protrusion extends (Plate I., fig. 3). The 
roughness of the surface of the shoot is determined by the number of papillae (Plate I., 
fig. 4). The core of the papilla appears to be mucilaginous, and on the addition of 
water it swells, bursting the cuticle with which it is covered. Owing to the similarity 
of these papillae to those found in leaves, and described by Haberlandt (l) as light- 
sense organs, an attempt was made to test their optical properties, Graham (2). In 
this, methods essentially those of Haberlandt were used. A piece of the young shoot 
about 1 cm. in length was divided into two. One half was then laid upon a glass slide, 
while the epidermis was freed from the adjacent tissues by gently scraping with a 
scalpel. The epidermis was now mounted upside down on a cover-glass, which was 
then used to roof a moist chamber about 4 mm. deep. The preparation was then 
placed on the stage of a microscope, extraneous light being shut out by enclosing the 
stage and tube in a black hood. The sub-stage iris was partly closed, when the whole 
field appeared darkened, with the exception of certain lighted areas corresponding in 
position with the papillae (Plate I., fig. 5). Further, the image of an object, either 
stationary or in motion, placed before the microscope, could be clearly seen in the light 


centres. It was found to be possible to obtain photographs both of the light discs and 
of the images in them of objects held before the mirror (Plate L, fig. 6). The occur- 
rence of these organs in stems which function to a great extent as leaves is almost to be 
expected, and the advantage of the erect light position to the desert forms of the 
Ephedrese is obvious from an oecological point of view. Any movement of the stem 
from the normal light position will be followed by a transference of the lighted area to 
another portion of the cytoplasm of the back wall of the epidermal cell. This trans- 
ference, accompanied as it must be by a variation in the intensity of the illumination, 
will act as a stimulus resulting in growth movements, whereby the former light position 
will be returned to. 

Stomata in single or double rows occur in the furrows between the ridges. Each 
stoma is sunk in a compartment below the surface of the epidermis (Plate J., fig. 7). 
The entrance to the ante-chamber is more or less constricted by the encroachment of 
the walls of the four surrounding cells (Plate I., fig. 8). The encroaching wall contains 
a central core of mucilage, and in E. viridis, where the material had been treated with 
water, the mucilage had swollen, causing the opposite walls of the chamber to meet. 
This suggests an adaptation for either narrowing or closing the chamber, and thus 
limiting transpiration. This is the probable explanation of the closing of the stomatal 
chambers by resinous masses, referred to by Volckens, and instanced by Stapf (3). 
The surface shape of the aperture depends upon the degree of encroachment of the 
projecting walls of the ante-chamber. The guard cells are without the usual ridges of 
entrance and of exit, as described by De Bary * (4). 

The Cortex is well developed, consisting of chlorenchyma and stereom. The 
chlorenchyma is differentiated into an outer palisade cortex of radially elongated cells, 
and an inner spongy cortex of polygonal cells, both having a well-developed aerating 
system, especially the latter (Plate I., fig. 9). Starch grains occur all through the 
cortex along with crystals of calcium oxalate, the latter being more abundant in the 
endocortex, where crystal sacs may occur. Calcium oxalate crystals also occur in the 
cell walls. In some species a very large proportion of the cortical cells contain tannin, 
either in the form of mucilage or in the form of small globular masses (Plate I., fig. 2). 
A sinuous layer of tangentially elongated, closely packed cells, containing large starch 
grains surrounding the stele, constitutes an endodermis. The innermost layer of the 
cortex abuts on to this layer at right angles. 

In the stem occur typically four series of strengthening fibres. The walls are at 
first of condensed cellulose, colouring blue only after prolonged treatment with iodine 
and sulphuric acid ; afterwards the walls become lignified, and are coloured red by 
phloroglucin and hydrochloric acid. Excluding the perimedullary stereom system, 
reference to which will be made later, the remaining three series of fibres are arranged 
on a girder principle, that is, a broad flange at either end, the hypodermal and 

* The outer and inner walls of the guard cells are lignified from an early date, a condition found in many of the 


pericyclic stereoms, linked radially by an incomplete mesocortical web (Plate II. fig. 10). 
In all the specimens examined the contours of the hypodermal groups of fibres were 
fairly uniform, being a flange or triangle with a longer or shorter portion projecting 
centripetal ly. The mesocortical and the pericyclic groups of fibres vary considerably, 
in some cases inversely. The mesocortical web consists of isolated fibres or groups of 
fibres, usually numerically less than the hypodermal or pericyclic groups. In some 
cases the mesocortical groups are poorly developed, and in a few cases are almost 
suppressed ; when this occurs the hypodermal is usually well developed. The single 
mesocortical fibres resemble those of Welwitschia in having calcium oxalate deposited 
in their external membranes (Plate II., fig. 10). This is also the case in the hypodermal 
fibres of E. Helvetica. The pericyclic stereom appears as a series of hard bast crescents 
or aggregates associated with the primary vascular bundles, one large or a few smaller 
ones to each (Plate II. , fig. 11). Between adjacent crescents occasionally a series of 
isolated fibres occur. 

A perimedullary stereom in the form of a discontinuous ring of sclerenchyma occurs 
in some species (Plate II., fig. 12) ; in others it is represented by a few isolated fibres or 
nests at intervals, while it may be absent altogether. 

Vascular System. — The vascular system resembles that found in Equisetum, with 
these differences, that in this case the leaf-trace is paired, that it is continued in the 
cauline part of its course through two internodes, and that in certain species an 
accessory bundle accompanies each leaf-trace through part of its course, while some 
species show a fusion of the leaf-traces in part of their course. With these exceptions 
the bundles in each describe a similar course, alternating in each internode and forming 
vascular networks at the node. The leaves, reduced to mere scales, are situated in twos 
or threes at each node. In the former case the arrangement is opposite and decussate, 
in the latter the whorls alternate. When there are two leaves at the node, the normal 
shoot shows in cross section a system of six, eight, or ten collateral endarch bundles 
surrounding a large pith. The structure of the primary bundle is as follows 
(Strasburger, 5). The whole is surrounded by a parenchyma sheath, some of the cells 
of which may contain chlorophyll. The primary phloem consists of cambiform 
parenchyma elements and narrow sieve-tubes with oblique plates. The primary xylem 
consists of spiral elements and bordered pitted elements with parenchyma packing 
tissue. The bordered pitted elements include both tracheids and vasa, the latter 
having their end walls steeply inclined and perforated by one or two rows of slightly 
bordered holes (Von Mohl, 6). Between the bundles medullary rays occur, but soon 
the bundles are linked up by the completion of a cambium, ring. Projecting from the 
xylem elements of the system in some species isodiametric lignified cells with reticulate 
thickening occur, and these at an early stage link up the adjacent bundles (Plate II., 
fig. 11). These cells are the first products of the activity of the cambium. Extending 
from the xylem to the medullary rays, as they do, they facilitate a rapid lateral 
distribution of crude sap to the chlorenchyma. 


The leaf-trace, composed of two bundles, passes down the stem vertically through 
two internodes. At the third node from that of their entrance to the stem, the bundles 
fuse right and left with the corresponding bundles of the traces emerging at this node 
(Plate II., fig. 13). Thus the first internode from an apical bud which has the vascular 
tissue developed shows in cross section four bundles, while in the second and subsequent 
internodes eight bundles are seen, two pairs of large bundles forming the traces of the 
leaves at the second node above, alternating with two pairs of small traces supplying 
the leaves at the first node above. A similar system is found in those species which 
have three leaves at each node ; the extra trace pairs in this case bring the number of 
bundles in each internode up to twelve (Plate II., fig. 15). In the internode of species 
whose phyllotaxis changes from two to three, the internode succeeding the node at 
which the alterations occur shows ten bundles. 

Immediately above each node occurs a region of two or three layers of compressed 
living cells with thick cellulose walls forming a dehiscence layer (Plate III., fig. 16). 
This layer, with in some cases a small area of meristematic tissue in the node itself, 
arises from the remains of the meristematic tissue from which elongation of the internode 
took place (Plate III., fig. 17). The layer partially cuts through the vascular tissues, 
the connection of the xylem elements being kept up by means of short reticulate 
tracheids. In the cortex the layer also appears, and there is a slight ring-like con- 
striction of the surface of the stem corresponding to it in position. The epidermal 
cells over the constriction are nearly isodiametric. OwiDg to the stoppage of the 
stereom systems on approaching this region, the internode is very brittle and readily 
breaks across at this point. While passing through the area of meristematic tissue the 
vascular bundles lose their accompanying lateral lignified flanges. 

At the node the stelar elements are concentrated into two vascular crescents, 
through the intercalation of linking tracheids. Spaces are left between the crescents 
corresponding with the position of the leaves at the node (Plate III., fig. 18). Each 
crescent is composed of the trace bundles from the leaf at the node above, flanked by 
a trace bundle from each leaf at the second node above (see Plate II., fig. 13). The 
leaf-traces of the leaves at the node in question link on to the crescent before pass- 
ing out to the leaves at the node. Just below the emergence of these leaf-traces the 
vascular supply for the axillary bud originates as a twig from each leaf-trace (see fig. 13). 

Lateral Branch. — Each of the bud- traces bifurcates to form the leaf- traces of the 
first pair of leaves on the side branch.. At the first node of the lateral branch 
vascular crescents are formed. The traces from the leaves at the two next higher 
nodes unite in pairs during their course in the second internode of the branch. 
Hence, apparently only two instead of the normal four bundles take part in the forma- 
tion of each vascular crescent at the first node (Plate III., fig. 19). 

Accessory Bundles (Strasburger's "complementary bundles " (7) ). — Originating 
from the centre of each vascular crescent, an accessory bundle accompanies each leaf- 
trace pair in the upper half of its course (Plate II., fig. 14). So far, only in one instance 

208 MR R. J. D. GRAHAM ON 

has the accessory bundle been found accompanying the leaf-trace in the lower half of 
its course. The accessory bundle, on reaching the node above that at which it 
originates, bifurcates at the level of the vascular crescents. Each half of the accessory 
then passes out as an additional pair of traces to the axillary bud (Plate III., fig. 20). 
The bud supply is hence augmented by two traces ; thus in cross section with the first 
shortened internode of the side branch six bundles appear arranged, two adaxially and 
four abaxially. Before the bud-traces pass out to the first leaves of the side branch, 
the accessory bud-traces fuse with them (see fig. 20). Thus the accessory bud supply 
is shared by the two leaves. In E. altissima, E. intermedia from Kew, in E. altissima 
from Algiers and from Montpellier, as also in E. distachya, elegantissima, lanceolata 
from Edinburgh, the leaves of the first node of the lateral branch are in many cases 
suppressed. This feature has been occasionally observed in other species. In correla- 
tion with the absence of leaves, no splitting occurs in the bud supply, while the 
accessory, when present, ends in the vascular crescents of the main axis. 

Concrescent Bundles.— This system is derived from the eight-bundle or normal 
type by the fusion of the adjacent leaf- traces of the opposite leaf- trace pairs in the 
second half of their course (Plate III., fig. 21, i.). At the end of its course the con- 
crescent trace splits and the two portions link on right and left with the traces on 
either side (fig. 21, ii.). Thus an incomplete ring is formed, interrupted only between 
the non-concrescent leaf-traces. At the node through the formation of the girdle of 
tracheids these non-concrescent traces become joined up (fig. 21, iii.). Thus, through 
the ultimate splitting of the concrescent bundle vascular crescents are formed composed 
of the same constituents as the vascular crescents in the normal type (fig. 21, iv.). 

Occasionally another type of concrescence occurs through fusion of the leaf-traces 
in the first half of their course. Exceptionally, concrescence takes place in all the leaf- 
traces, and the number of the bundles is reduced to four. The leaf-traces in the leaves 
appear to be very close together in this last case. 

When there are three leaves at the node the typical vascular system is still retained. 
If the leaf-traces be accompanied by an accessory, the number of bundles seen in 
section is fifteen (Plate III., fig. 22) ; while if concrescence occur in the latter half of 
the course of the bundle, the number of bundles seen is nine (Plate III., fig. 23). 

These variations due to intercalation and concrescence of bundles may take place at 
different levels in the course of the trace, and need not be synchronous on the two 
sides. Hence in different internodes of the same individual four, five, six, seven, eight, 
nine, and ten bundles may be met with. Similar variations occur in those species 
which have three leaves at the node. 

The Pith consists of large and small cells, whose walls from an early stage are 
lignified, except around the intercellular spaces. Lignification is centripetal, being 
found in the fourth internode of E. fragilis, v. campylopoda. The perimedullary 
stereom (Plate II., fig. 12) has already been noticed. Cells containing tannin mucilage 
occur in many species, though they may be entirely absent from others (Plate III., 


fig. 22). Certain of the pith cells in the neighbourhood of the protoxylems show 
greenish-coloured contents, which seem to be chlorophyll. The cells of the pith have 
pitting on all their walls, especially on the end walls (Plate III., fig. 23). 

Meristematic Tissues. — The stem apex, enclosed in several whorls of leaves, consists 
of a small conically-shaped mass of tissue. According to Dingler (8), growth is from a 
tetrahedral apical cell. In all the material examined the meristem has been found to 
be stratified. Different records in regard to the nature of the apex have, however, been 
made, and De Bary (9) mentions that both forms of apex occur in the same species. 
In the first internode after the apical the stelar elements have begun to differentiate, 
while in the second internode they are well developed. In this internode elongation 
of the stem has set in, but is more pronounced in the succeeding two or three inter- 
nodes. At first the whole internode is meristematic (Plate III., fig. 17), but later 
elongation comes from a meristem situated at the base of the internode. This basal 
meristem remains functional for some considerable time, eventually passing over into 
permanent tissue, with the exception of a layer of from two to three cells in thickness, 
situated just above the node. This layer functions as a dehiscence layer (Plate III., 
fig. 16). 

Part II. — Histology applied to the Determination of Species. 

Considerable attention has recently been paid to the use of the internal structure of 
the vegetative organs of plants in the diagnosis of orders, genera, and species. The 
genus Ephedra, consisting as it does of a uniform oecological group of species, seems to 
furnish a fair test of the value of such diagnostic characters. In his paper on Gnetacese 
and Coniferae, Bertrand (10) was of opinion that the histological features of stem and 
leaf were insufficient, and not thoroughly trustworthy for specific distinctions. This 
opinion is also held by Dr Stapf in his monograph, " Die Gattung der Ephedrege," in 
which he constantly emphasises the extreme variability of the histology of the vegeta- 
tive organs. While the tendency to variation in the histology is admitted, it has yet 
been possible, as a result of the above-mentioned notes, to divide the genus into sub- 
generic groups. Further, the additional evidence furnished from the internal structure, 
when taken with the ordinary external morphological characters, greatly facilitates 
diagnosis. The idea with which this part of the work was undertaken was to work from 
some definite level, and thereafter to include such other features from other parts of 
the stem as might seem to be of value. The second elongated internode from an 
apical bud was taken to start with, as being the youngest in which the complete 
primary stelar structure is attained. A general division of the group can be made, 
basing the evidence on the vascular supply of this internode. The number of bundles 
is very constant in some species (eight in foliata, Helvetica, procera, distachya ; ten 
in viridis); while in a few there is a slight tendency to vary (nebrodensis and 
gerardiana), and in even fewer variations actually occur (trifurca). These variations 
are never of such an extent that they interfere with the value of the classification, for 

210 MR R. J. D. GRAHAM ON 

they never interfere with the average number of bundles, and hence the type can 
always be determined. Two other important features for distinction of species are 
mentioned by Fritsch (ll). These are the occurrence of isolated groups of 
sclerenchyma in pith and cortex, and the distribution of secretory organs. Both of 
these can be applied to the genus Ephedra. The development of the mesocortical 
stereom web, as already mentioned, is much better in some species than in others. A 
division can, therefore, be made into species which have a well-developed mesocortical 
stereom {foliata, distachya, viridis), and those in which the stereom is poorly developed 
{trifurca, fragilis, Helvetica, procera). Again, the extent to which the perimedullary 
stereom is developed furnishes a basis for another threefold division. The perimedullary 
stereom is represented in two stages of development — as a discontinuous ring (nebro- 
densis), as a series of isolated fibres {trifurca, foliata, distachya, viridis), or as absent 
altogether {fragilis, Helvetica, gerardiana, procera). The presence of tannin sacs in the 
pith of many species {Helvetica, distachya, procera) furnishes another basis for division. 
In determining the value of varieties, histology plays an important part. Thus 
E. nebrodensis differs from E. procera, not only in the former having a roughened 
epidermis while the latter is smooth, but E. nebrodensis has ten vascular bundles, while 
E. procera has only eight. A striking instance occurred in regard to three specimens 
received from Edinburgh. The names borne by the plants were apparently synonyms 
for E. altissima. Their identity was established by a study of their histology, and by 
a subsequent examination of the plants with regard to external morphology the 
identity was confirmed. 


1. E. trifurca (Torb). Primary bundles, 12 or 15. Mesocortical stereom web poorly developed. 

Perimedullary stereom isolated fibres. No pith tannin sacs. 

2. E. foliata (C. A. Meyer). Primary bundles, 8. Mesocortical stereom web well developed. Peri- 

medullary stereom isolated fibres. 

3. E. fragilis (Desf.). Primary bundles, 8. Mesocortical web poorly developed. Perimedullary 

stereom absent. No pith tannin sacs. 

4. E. Helvetica (C. A. Meyer). Primary bundles, 8. Mesocortical stereom web poorly developed. 

Perimedullary stereom absent. Pith tannin sacs. 

5. E. distachya (Lin.). Primary bundles, 8. Mesocortical stereom web well developed. Peri- 

medullary stereom isolated fibres. Pith tannin sacs. 

6. E. gerardiana (Wall). Primary bundles, 10. Perimedullary stereom absent. Pith tannin sacs. 

7. E. nebrodensis (Tineo). Primary bundles, 10 or 15. Perimedullary stereom discontinuous ring. 

No pith tannin sacs. 

8. E. procera (C. A. Meyer). Primary bundles, 8. Mesocortical stereom web poorly developed. 

Perimedullary stereom absent. No pith tannin sacs. 

9. E. viridis (Coville). Primary bundles, 10. Mesocortical stereom web well developed. Peri- 

medullary stereom isolated fibres. Pith tannin sacs. 

In conclusion, I wish to thank Mr R. A. Robertson, at whose suggestion the work 
was undertaken, for his great help throughout its course. 



Plate I. 

Fig. 1. Transverse section of E. Helve ica, showing epidermal cells with thickened external wall. 
(a) Cuticle ; (b) calcium oxalate layer ; (c) cellulose stratum, x 360. 

Fig. 2. Transverse section of E. fragilis (Algiers), (a) Division of epidermal cell into two ; (6) division 
of one of the daughter-cells by a wall at right angles to the first division wall ; (c) cortical tannin sacs. 

Fig. 3. Transverse section of E. viridis, showing large epidermal cells over ridge, one having a papilla 
on the external wall, (a) Cuticle; (b) calcium oxalate layer ; (c) cellulose stratum. x 360. 

Fig. 4. Surface view of E. viridis, showing papillae and stomata in alternating rows. x 42. 

Fig. 5. Photomicrograph of preparation of the epidermis of E. altissima, showing position of light spots. 
x 160. 

Fig. 6. Photomicrograph as above of E. altissima, showing image of St Andrew's cross in each light 
spot. x 160. 

Fig. 7. Transverse section of E. nebrudensis, showing sunk stoma, guard cells, respiratory chamber, 
ante-chamber, (a) Mucilaginous core of wall of cells forming ante-chamber. x 360. 

Fig. 8. Longitudinal section of E. distachya, showing stoma, respiratory chamber, imte-chamber. 
(a) Mucilage core of cell wall. x 360. 

Fig. 9. Transverse section of E. nebrodensis, showing hypodermal stereoin, ami! differentiation of cortex 
into palisade and spongy cortical cells containing starch grains and crystals of calcium oxalate. x 160. 

Plate II. 

Fig. 10. Transverse section of E. distachya, showing stereom systems well developed, (a) Welwitschia- 
like fibres ; (b) division of epidermal cell. x 160. 

Fig. 11. Transverse section of E. distachya, showing pericyclic stereom crescent and vascular flange 
projecting from xylem. (a) Pericyclic stereom crescent ; (b) vascular flange. x 160. 

Fig. 12. Transverse section of E. nebrodensis, showing perimedullary stereom (a). x 160. 

Fig. 13. Diagram of eight or normal bundle system, showing the entrance of the trace at a 1 , the parallel 
course of the bundles, the behaviour of the trace at the first node (a"), the continatiou through a second 
internode and the linking on of the traces to the bundles passing out at the third node (a 111 ). Vascular 
crescents (a) with their component parts are diagrammatically indicated. The bud-trace (b) is indicated in 
position below the passing out of the trace at a 1 . The splitting of the bud-trace is indicated. 

Fig. 14. Diagram of the ten or accessory bundle system. Explanation as in last. The accessory (c) is 
indicated splitting in the region of the vascular crescent at a i , where it passes out to form an extra pair of 

Fig. 15. Transverse section of E. trifurca, showing twelve bundles due to the presence of three leaves 
at the node. x 42. 

Plate III. 

Fig. 16. Longitudinal section of E. nebrodensis, showing dehiscence layer. x 42. 

Fig. 17. Longitudinal section of apex of E. foliata, showing development of young internode. x 42. 

Fig. 18. Transverse section of E. fragilis, v. campylopoda, showing vascular crescents with accessory 
splitting in between. x 42. 

Fig. 19. Diagram of junction of lateral branch with main stem, showing the bifurcation of the bud- 
trace (b) and its course into the leaves ; the fusion of the traces of the higher leaves of the lateral branch 
before their arrival at the first node (a), thereby making the crescent have the appearance of being made up 
of only two bundles. 

Fig. 20. Diagram of the junction of lateral branch in the accessory system. The main features are as 
above, while the splitting of the accessory bundle to augment the bud-supply is indicated (c) ; also the fusion 


of the twigs of the accessory with the normal bud-trace before they pass to the leaves is shown (/). The 
accessory is represented as being absent from the first two internodes of the side branch. 

Fig. 21. Diagram of the behaviour of the concrescent bundle at the end of its course, (i.) represents 
the six bundles after the disappearance of the lateral flanges; (ii.) represents the concrescent bundle linking 
on to the traces on either side ; (iii.) represents the concrescent bundle as split, while the non-concrescent 
bundles have become united by the formation of intercalary tracheids ; (iv.) represents the - true vascular 
crescents formed equally of the traces from the first node above, flanked on either side by half the con- 
crescent bundle. 

Fig. 22. Transverse section of E. trifurca, showing fifteen bundles due to three leaves at the node, 
each leaf-trace pair being accompanied by an accessory. x 42. 

Fig. 23. Transverse section of E. elegantissirna, showing nine bundles due to three leaves at each node, 
the leaf-traces being concrescent in the second half of their course, x 42. 


(1) Haberlandt, G., Die Lichtsinnesorgane der Laubbldtter, Leipzig, 1905. 

(2) Graham, R. J. D., " On the Light-sense Organs in Ephedrese," Trans. Scots. Mic. Soc, March 1907. 

(3) Stapf, O., "Die Gattung der Ephedreaa," Denkschr. Akad. Wissensch. Wien, 1889, Bd. lvi. 

(4) De Bary, A., Comjiartitive Anatomy of the Phanerogams and Ferns, Oxford, 1884, p. 35. 

(5) Strasburger, E., " Ueber den Bau und die Verrichtungen der Leitungsbahnen in den Pflanzen," Hist. 

Beit., 1891, Heft iii., p. 140. 

(6) Von Mohl, H., " Ueber den Bau der grossen getiipfelten Rohren von Ephedra," Linnea, 1831, Bd. iv. 

p. 593. 

(7) Strasburger, E., Die Coniferen u. die Gnetaceen, Jena, 1872, p. 115. 

(8) Dingler, H. "Ueber das Scheitelwachsthum des Gymnospermen-Stammes, " Bot. Zeit., 1882, Bd. si., 

p. 796. 

(9) De Bary, A., I.e., p. 14. 

(10) Bertrand, "Anatomie des Gnetacees et des Coniferes," Ann. d. Sc. Nat., 1874, Ser. v., Bot., torn, xx., 

p. 24. 

(11) Fritsch, F. E., "The Use of Anatomical Characters for Systematic Purposes," New. Phyt., 1903, vol. ii., 

p. 177. 

Trans. Roy. Soc. Edinburgh. 

Vol. XLVI. 

R. J. D. Graham : " Histology of the Ephedrefe." — Plate I. 

Fig. 1. 

Fig. 3. 

Fig. 4. 

Fig. 5. 

Fig. 6. 

Fig. 7. 

Fig. 8. 

Fig. 9. 

Trans. Roy. Sue. Edinburgh. 

Vol. XLVL 

R. J. D. Graham : " Histology of the Ephedrete." — Plate II. 

Fig. 10. 

Fig. 11. 


■ * ■ :-. 


Fig. 12. 

Fig. 15. 



I j 






Fig. 13. 

Fig. 14. 

Trans. Roy. Soe. Edinburgh. 

A t ol. XLVI. 

R. J. D. Graham: "Histology of the Ephedrese." — Tlate III. 

•" i0f7''i ; '«f it" ? 

Fig. 16. 

Fig. 18. 

Fig. i; 

Fig. 21. 

Fig. 19. 

Fig. 20. 

. .%::. ,~ ,_ 

Fig. 23. 

Fig 22. 

,^SH M^ 


TIi^Transactions of the Royal Society of Edinburgh will in future be Sold 

at the following reduced Prices :- 



Price to the 

Price to 

• V 

Price to the 

Price to 






£0 11 


£0 9 6 

XXX IX. Part 1. 

£1 10 





Part 2. 


14 6 




Part 3. 

2 3 

1 11 




Part 4. 






XL. Part 1. 







„ Part 2. 

1 12 6 

1 5 6 





„ Part 3. 

1 6 

19 6 




„ Part 4. 




1 5 

1 1 

XLI. Part 1. 

1 1 

15 9 


1 11 

1 6 

„ Part 2. 

1 9 6 

1 2 

XX. Part 1. 



„ Part 3. 

•2 5 

1 13 6 

XXII. Part 2. 


7 6 

„ Part 3. 

1 5 

1 1 

XXVII. Part 1. 



Part 2. 


4 6 

Part 4. 



XXVIII. Part 1. 

1 5 

1 1 

Part 2. 

1 5 

1 1 

Part 3. 


13 6 

XXIX. Part 1. 

1 12 

1 6 

Part 2. 



XXX. Part 1. 

1 12 


Part 2. 



Part 3. 



Part 4. 



5 8 


4 4 

3 3 

XXXII. Part 1. 



Part 2. 


13 6 

Part 3. 

2 10 

1 17 6 

Part 4. 



XXXIII. Part 1. 

1 1 


Part 2. 

2 2 

1 11 


Part 3. 


9 6 


2 2 

1 11 

XXXV.*Part 1. 

2 2 

1 11 

Part 2. 

1 11 

1 3 6 

Part 3. 

2 2 

1 11 

Part 4. 

1 1 


XXXVI. Part 1. 

1 1 


Part 2. 

1 16 


1 7 6 

Fart 3. 



XXXVII. Parti. 

1 14 


1 5 6 

Part 2. 

1 1 


Part 3. 



Part 4. 



5 8 



1 10 

Part 2. 

1 5 


Part 3. 

1 10 

1 3 

Part 4. 



5 8 

* Vol. XXXV., and those which follow, may be had in Numbers, each Number containing 

a complete Paper. 

January 1906. — Volumes or parts of volumes not mentioned in the above list are not for the present on sale 
public. Fellows or others who may specially desire to obtain them must apply direct to the Society. As the 
reprints from time to time parts of its publications which have become scarce, the absolute correctness of tl 
cannot be guaranteed beyond this date. 






No. Page 

VII I. On the Histology of the Ephedreas, toith Special Reference to the Value of Histology for 

Systematic Purposes. By R. J. D. Graham, M.A., B.Sc. (With Three Plates), . 203 

{Issued separately February 20, 1908.) 

IX. On the Fossil Osmundacece. Part II. By R. Kidston, F.R.S.L. & E., F.G.S., Foreign Mem. 
K. Mineral. Gesell. zu St Petersburg ; and D. T. Gwynne-Vaughan, M.A., F.L.S., 
Lecturer in Botany, Birkbeck College, London. (Plates I.-IV.), . . .213 

(Issued separately March 27, 1908.) 

X. The Medusce of the Scottish National Antarctic Expedition. By Edward T. Browne, 

Zoological Research Laboratory, University College, London. (With Two Plates), . 233 

(Issued separately May 1, 1908.) 

XL On q-Functions and a certain Difference Operator. By the Rev. F. H. Jackson, M.A., . 253 

(Issued separately April 21, 1908.) 

XII. The Evolution of the Eyebrow Region of the Forehead, with Special Reference to the Excessive 
Supraorbital Development in the Neanderthal Race. By Professor D. J. Cunningham, 
F.R.S. (With Three Plates), ........ 283 

(Issued separately June 16, 1908.) 

XIII. On the Age of the Old Red Sandstone of Shetland. By John S. Flett, M.A., D.Sc, . 313 

(Issued separately July 8, 1908.) 

XIV. On Fossil Fish-remains collected by J. S. Flett, M.A., D.Sc, from the Old Red Sandstone of 

Shetland. By Ramsay H. Tkaquaik, M.D., LL.D., F.R.S. (With Two Plates), . 321 

(Issued separately July 4, 1908.) 

V V. On the Lamellibmn, ?. , ni j Gasteropod Fauna found in the Millstone Grit of Scotland. By 

Wheelton t >, Ai.j..., T3 #S-> F.R.C.S., F.G.S. (With Two Plates), . . .331 

(Issuea ^.-rately September 2, 1908.) 
XVI. On a New Species of Dineuron and of ., teris f rom p e ttycur, Fife. By R. Kidston, 

LL.D., F.R.S., F.R.S.E. (With One PL . , _ . .361 

(Issued separately A' ..'5, i 

XVII. The Craniology, Racial Affinities, and Descent of the Aborigines of /'&» ■- p, v Principal 

Sir William Turner, K.C.B., D.C.L., F.R.S. (With Three Plates; , 365 

(Issued separately October 16, 1908.) 

l/ XVIII. Les Holothuries de V Expedition Avtarctique Nationale ficossaisr. Par Clement Vaney, 
Maitre de conferences de Zoologie a la Faculto des Sciences de Lyon. (Avec cinq 
planches), .......... 405 

(Issued separately October 15, 1908.) 




Price Twenty-five Shillings and Eightpence. 

( 213 ) 

IX. — On the Fossil Osmundacese. By R. Kidston, F.R.S. L. & E., F.G.S., 
Foreign Mem. K. Mineral. Gesell. zu St Petersburg ; and D. T. Gwynne- 
Vaughan, M.A., F.L.S., Lecturer in Botany, Birkbeck College, London. 
(Plates I. -IV.) 

(MS. received January 6, 1908. Read same date. Issued separately March 27, 1908.) 


Zalesskya gracilis, Eichwald, sp., and Zalesskya diploxylon, Kidston 

and Gwynne-Vaughan, n.sp. 

While the first part of this paper was in course of preparation a search was made 
through various palseobotanical publications in the hope of meeting with records of 
Osmundaceous fossils that had hitherto escaped recognition as such. In so doing our 
attention was at once attracted by the descriptions and figures given by Eichwald in 
his Lethwa Rossica of some very fine fossils which were held by him to represent 
the stems of arborescent ferns. Those in which we were particularly interested were 
Chelepteris gracilis, Eichwald,! Sphallopteris Schlechtendalii, Eichwald,| Bathypteris 
rhomboidea, Eichwald,§ sp., and Anomorrhaa Fischeri, Eichwald.|| The figures and 
descriptions of these plants were quite inadequate for the determination of their true 
affinities ; but it must be remembered that at the time when these descriptions were 
written by Eichwald many structural characters which are now regarded as of 
paramount importance in determining affinities were not recognised as such, even by 
botanists. To enable us if possible to clear up these difficulties, we wrote to Mons. 
Michel Zalessky, geologist on the staff of the Comite geologique, St Petersburg, 
asking if the types of these specimens were known to be in existence, and if any 
preparations had been made from them suitable for microscopical examination. 

On receiving our letter Mons. M. Zalessky instituted a most laborious search in 
several museums, and was eventually successful in finding the types of the four genera 
already mentioned in the Museum of the Institute of Mines, St Petersburg. No 
sections for microscopical examination had ever been prepared from any of the stems, 
though the microscopical structure of the petiole base of Sphallopteris Schlechtendalii 
had been imperfectly described by Eichwald,H and that of Bathypteris rhomboidalis 
by Schmalhausen. ## 

* Part I., Trans. Roy. Hoc. Edin., vol. xlv., part iii. (No. 27), pp. 759-780, pis. i.-vi., 1907. 

t Letheea Rossica, vol. i., p. 98, pi. iii., figs. 4, 5, and 6, 1860. 

| Ibid., vol. L, p. 93, pi. iii. figs. 1-2, pi. xx. figs. 2-5. 

§ Ibid., vol. i., p. 96, pi. iv. figs. 1-2 ( = Tubicaulis rhomboidalis, Kutorga (pars), Verhandl. d. Miner. Gesell. zu 
St Petersburg, pi. i. fig. 6, 1844). 

|| Ibid., vol. i., p. 102, pi. iv. figs. 3-4. 

IT Ibid., pi. xx. figs. 2-5. 

** " Die Pflanzenreste d. Artinskischen und Permisehen Ablagerungen," etc., Mem. du Comite' ge'ol. (St Petersburg), 
vol. ii., No. 4, pp. 9 and 36, pi. iii. figs. 6-7, 1887. 



The references to the internal structure of Chelopteris gracilis and Sphallopteris 
Schlechtendalii, embodied in Eichwald's descriptions of these species, only refer to 
characters observable by the hand-lens. 

With the object, therefore, of submitting the stems of Chelepteris, Sphallopteris, 
Anomorrhcea, and Bathypteris to a careful microscopical examination, the Director of 
the Comite geologique had portions cut from each of these stems, and most generously 
presented them to us for examination ; and we take this opportunity of expressing our 
indebtedness and sincere thanks to Mons. Th. Tschernyschew for supplying us with 
the necessary material for our examination of these genera, and to Mons. Michel 
Zalessky for the willing labour bestowed on searching for the specimens. To Mons. 
Zalessky we are further indebted for excellent photographs of the external surface of 
the specimens, which were taken by Mons. Koch. 

Although in the present paper only Chelepteris gracilis, Eichwald (for which, for 
reasons stated later, we propose the genus Zalessky a), and a new species of the same 
genus are described, it is desirable, before proceeding further, to consider the relationship 
and distinctive characters of Chelepteris, Spliallopteris, Bathypteris, and Anomorrhwa, 
all of which are placed by Corda # and Eichwald t in the Protopteridese, the chief 
character of which is the horse- shoe shaped vascular bundle of the petioles. 

Chelepteris, Corda. 

1845. Chelepteris, Corda, Flora d. Vurwelt, p. 76. 

Stem arborescent, round, erect ; exposed cicatrices of leaves supported on elevated 
cushions, spirally arranged and ornamented with adventitious rootlets. Cicatrices 
oblong or ovate, with a single median crescent- shaped vascular bundle. 

This genus was founded for the reception of three fern stems which were described 
by Schimper and Mougeot under the names of Caulopteris Voltzi,\ Caulopteris 
micropeltis,^ and Caulopteris Lesangeana. \\ 

Eichwald, who adopted this genus for his Chelepteris gracilis, thus describes it: — 

Chelepteris (Corda), Eichwald. 
1860. Chelepteris, Eichwald, Letheea Rossica, vol. i., p. 98. 

Stem arborescent, straight, cylindrical, formed of petiole bases, very close and 
cylindrical, disposed vertically in spiral series and showing on their fractured end a 
single semilunar vascular bundle with rarely occurring adventitious roots in the 
interstices of the petiole bases ; the ligneous body (stele) is complete and its ivalls 
are stout. 

* Flora d. Vox well, p. lb. \ Loc. at., p. 92. 

% Monog. d, pluntes foss. du grea biyarre de la chuine des Vosges, p. 65, pi. xxx., pi. xxxi. figs. 1-2, 1844. 

§ Ibid., p. 67, pi. xxxi. fig, :;. || Ibid,, p. 67, pi. xxxii. 


In addition to a few structural features added by Eichwald, the leaves are said to 
be very close. 

Sphallopteris, Corda (emend.). 

1845. Sphalmopteris, Corda, Flora der Vorwelt, p. 76. 

Stem arborescent, round, external cushions elevated, spirally arranged (5/8). 
Cicatrices showing the impression of a centred depressed simple horse-shoe shaped 
vascular bundle. 

The type of Corda's Sphallopteris is the stem which Brongniart referred to his 
Anomopteris Mougeotii* and which is figured in the Hist. d. veget. foss , vol. i., pi. lxxx.f 

Brongniart believed this fern stem to have borne the fronds he named Anomopteris 
Mougeotii, | on account of their large size and being found in the same bed as the stem. 
They were not, however, found in organic union, hence the uncertainty of the stem and 
fronds belonging to the same species ; and this consideration induced Corda to place the 
stem in a separate genus. 

It is very interesting to notice that traces of a semilunar vascular bundle were seen 
on some of the broken-over petioles, which are compared to those of Osmunda regalis 
by Brongniart. 

Eichwald thus describes the genus : — 

Sphallopteris (Corda), Eichwald (emend.). 
1860. Sphallopteris, Eichwald, Lethsea Rossica, vol. i., p. 92. 

The exterior of the stem is composed of a cortex form,ed of adventitious rootlets 
and petiole bases or cushions disposed in regular spiral series, and which are very 
prominent and very large, and show in transverse section cicatrices in the form of a 
horse-shoe in the middle or on their surfaces. 

" The cushions are triangular or almost rhomboidal, the two lateral edges are always 
acute, the inferior or anterior margin is also acute, but the superior or posterior margin 
is rounded, rarely pointed, and is not raised up above the surface of the stem, as the 
inferior ; it loses itself likewise in the leaf ; the horse-shoe shaped vascular bundle has 
its two extremities bent in as a hook. 

" The interpulvinar area is entirely riddled with vascular bundles, which form some- 
times vertical rows, or which are sometimes disposed horizontally, leaving large furrows 
as the traces of their former existence. 

" The woody body (cylindrus ligneus) is small, narrow, and occupies the axis of the 
stem ; the vascular bundles which compose it diverge in all directions and ascend 
obliquely" (the leaf-traces). 

* " Essai d'une Flore d\\ gres bigarre," Ann, d. Sc. Nat, Dec. 1828, p. 4. 

t P. 261, 1831 or 1832. 

| Loc. cit., pis. lxxix. and lxxxi. 


Eichwald, believing that his stem was generically identical with that described by 
Brongniaut, and for which Corda founded the genus Sphallopteris, adds to the 
original characters some which are derived from the internal structure of the stem. 

One of the chief distinctive differences which separates Chelepteris from Spthallo- 
pteris, as pointed out by Eichwald, is the smaller vascular axis of the latter when 
compared with that of the former. 

Bathypteris, Eichwald. 

1860. Bathypteris, Eichwald, Lethsea Rossica, vol. i., p. 96.] 

Stem large, simple, almost cylindrical, swollen at the middle, and contracted 
towards the extremities ; the surface is composed of contiguous petiolar bases, discs 
( = broken-over surface of petioles) much sunk in, almost round or somewhat squared, 
provided in the centre with a small semicircular cicatrice ; the margins of the petiole 
bases only are prominent and riddled ivith holes or canals arising from the adven- 
titious roots ; the elongated leaves, closely placed beside each other, are fixed at the 
upper extremity of the stem. 

This genus is said to differ from Chelepteris and Sphallopteris in the petiole bases 
being scarcely visible at the surface of the stem, whereas in Chelepteris and Sphallo- 
pteris the stems possess prominent petiole bases. It will be seen as we proceed that 
this character depends on the state of preservation of the specimens. The stele of 
Bathypteris was not preserved. 

Anomorrhcea, Eichwald. 

I860. Anomorrhoea, Eichwald, Lethxa Rossica, vol. i., p. 102. 

Stem fairly large and composed of the bases of almost rhomboidal-tubidar petioles 
which are provided with many layers, ascending obliquely and placed quincuncially. 
They enlarge insensibly towards the extremity and are provided with a semicircular 
vascular bundle, sometimes in the form of a horse-shoe, of which the centre is hollow ; 
the interpetiolar interstices have no adventitious roots, by which the genus is princi- 
pally distinguished from the genera mentioned, above* 

The stele was not preserved. 

Eichwald was mistaken in supposing that Anomorrhoea did not possess adven- 
titious roots, for sections of the specimen show them to be present as in all the other 
three genera to which reference has been made. 

These four genera have been differently regarded by subsequent writers, but it 
must be borne in mind that probably all those authors who have referred to them had 
only the descriptions and figures for their guidance, for it is certain that no micro- 
scopical preparations were made from the specimens at the time they wrote, and it is 
also further highly probable that they were unable to examine the original types. 
* The genera referred to by Eichwald are Sphallopteris, Bathypteris, Chelepteris, and Desmia. 


Brongniart proposes for the fern stem which was first described by Eichwald 
under the name of Anomopteris Schlechtendali* and subsequently placed by him in 
Sphallopteris, Corda, the new genus Thamnopteris,^ on the ground that Eichwald's 
stem was not generically similar to the Anomopteris Mougeotii I which Corda had 
made the type of his genus Sphalmopteris ( = Sphallopteris, Eichwald). 

Brongniart says in regard to his genus Thamnopteris : " I place under this name a 
frutescent erect fern stem, but with persistent petioles arranged around a very slender 
stem, which M. Eichwald designates under the name of Anomopteris Schlechtendali, 
because it appeared to him to offer some analogy with the stem which I have referred 
to Anomopteris Mougeotii ;§ but this latter differs notably in its very slender stem 
surrounded by very numerous slender petioles, and which have evidently been persistent 
as those of our herbaceous ferns with ascending stems." || 

We have not seen the specimen of Sphallopteris Schlechtendalii to which Brongniart 
refers, but, judging from the specimen of Sphallopteris Schlechtendalii which Eichwald 
gives on PI. III., fig. 2, of the Lethsea Rossica, one would scarcely be inclined to include 
it in the same genus as Anomopteris Mougeotii, and in our future treatment of this 
fossil we shall adopt the name of Thamnopteris Schlechtendali for this stem. 

There are other considerations which advocate this course, and one of these is that 
a microscopical examination of the specimen of the Thamnopteris Schlechtendali figured 
by Eichwald in the Lethsea Rossica reveals the fact that the true outer surface of the 
stock is not preserved, but the stem was surrounded by a thick mantle of concrescent 
leaf-bases, and an unknown thickness of these concrescent leaf- bases has been removed 
from the stem. Probably the true appearance of the outer surface of the stock would be 
a felted mass of rootlets in which the remains of the petioles were embedded. In fact, 
this specimen and those we have examined of Zalesshya [Chelepteris, Eichwald non 
Corda), Anomopteris, and Bathypteris are all more or less incomplete or even partially 
decorticated as in Zalesshya, and none of them show the true outer surface of the stock, 
and they might well be described as all representing a " Knorria" condition. There- 
fore, if the figure given by Brongniart as Anomopteris Mougeotii shows its true outer 
surface (which it may do), a comparison has been made between structures which hold 
different positions in the stock. 

But a more weighty reason for the rejection of the genus Sphallopteris for 
Eichwald's plant is found in the circumstance that the structure of the type of the 
genus Sphallopteris (Anomopteris Mougeotii, Brongniart) is not known, and there is no 
warrant for inferring that it was similar to that of Eichwald's specimen. All that is 
known of the structure of Anomopteris Mougeotii is that the petioles seem to 
possess semilunar vascular bundles " as in the petiole of Osmunda, regalis,'' but this 

* Eicswald, Urwelt Eusslands, Heft ii., p. 180, pi. iv., figs. 3-5. 

t Brongniart, Tableau d. genres d. ve'ge't.foss., p. 35, 1849. 

I Brongniart, Hist. d. ve'ge't.foss., vol. i., pi. lxxx. 

§ The Cottce Mougeottii of Schimper and Mougeot, Plantesfoss. du gres bigarre', p. 69, pi. xxxiii. 

|| Tableau d. genres d. ve'ge't.foss., p. 35. 


character is of too general occurrence to be accepted as indicating generic identity 
between the fern stems in question. The genus Tliamnopteris, as it will be employed 
by us, when we come to describe the specimen in detail, will be one which derives its 
essentia] characters from the anatomy of the stem. We take this liberty of altering 
the generic characters on the strength of the fact that the stem we hope to describe, 
though not actually the individual to which Brongniart's remarks refer, is one of 
two which were figured and described by Eichwald as belonging to his Sphallopteris 

It is similar considerations which have prompted us to remove the Chelepteris 
gracilis, Eichwald, from that genus and to place it in the new genus Zalesskya. This 
stem has not only lost its mantle of leaf-bases, but is also partially decorticated, so its 
outer surface is unknown. The whole of the characters of the genus Chelepteris, Corda, 
are derived from external features, and possibly the specimens placed in Chelepteris 
by Corda may show their outer surface ; but be this as it may, the genus Zalesskya is 
founded on characters derived entirely from its interna] organisation — an organisation 
which one has no right to infer was possessed by the original types of the genus Chelepteris. 

It is difficult and probably useless to attempt to define the relationship of the plants 
placed in Chelepteris, Corda, to the Chelepteris (Zalesskya) gracilis of Eichwald ; but 
we are of opinion that they are probably generically distinct, though it is impossible 
to determine this point satisfactorily. 

The genera Chelepteris, Corda (non Eichwald), and Sphallop)teris, Corda (emend.), are 
placed by Brongniart alongside of Protopteris, which is compared with the Dicksonise, 
though he points out that the same type of leaf-trace occurs in the Osmundacese and 
other ferns. 

Though the genus Protopteris, Presl,* may be Cyatheaceous, we are clearly of opinion 
that the fern stems described by Eichwald under the names of Chelepteris and 
Sphallopteris (without expressing any opinion on the fossils placed in these genera by 
Corda) are Osmundaceous, and, along with Zalesskya and Anomorrhcea, must be 
removed from the Protopteridese. 

Schimper, in his Traite d. paleont. veget.,\ treats Thamnopteris, Anomorrhcea, 
Chelepteris, and Bathypteris as distinct genera, hutinZAtteYsHandbuchd.Pcdaeontologie]. 
he unites Anomorrhcea and Bathypteris with TJiamnopteris, remarking that there are 
no grounds for placing these fern stems in different genera as has been done, as the 
outer surfaces of all show the same characters ; and had not one the structure of the 
specimens to direct them, this appears to us as the wisest way of treating the fossils, 
though, as will be seen later on, we still keep these genera separate, but from an 
altogether different class of characters than those observable on their outer surfaces. 

We have thought it necessary to give this review of the literature of the subject, to 
show the opinions held by the various botanists who have written about these fossil fern 

* PKESL in Sternberg. Vers., vol. ii., fasc. vii.-viii., p. 169, 1838. 

+ Vol. i., pp. 701-3, 1869. 

I Schimi'ek in Zittel, Handh. d. Palaeont., II. Abth., Palaeophylologie, p. 145, 1879. 



stems ; but it must be remembered that probably none of these writers have had the 
opportunity of examining the specimens, and the published descriptions are quite in- 
adequate to enable one to give a true interpretation of their systematic position. 

As already stated, for the Chelepteris gracilis, Bichwald, we have been compelled to 
form a new genus, and separate it from those species originally placed in Chelepteris by 
Corda ; but as the other three genera, Thamnopteris, Anomorrhcea, and Baihypteris, 
only contain the type specimens, and these are the specimens investigated in this and 
the succeeding paper, we are fortunately enabled to retain the old names, though the 
characters on which we now define these genera are entirely derived from their internal 

Our specimens of Zalesskya, Thamnopteris, Anomorrhcea, and Baihypteris are from 
the Upper Permian of the Oural, and their geological position is seen in the annexed 
table, which has been kindly communicated to us by Mons. Zalessky : # — 


Division du 

permien et 


en Eussie. 

Le permien d'Oural, gouv. Oufa et 









P 3 


Marnes bigarres avec les conchiferes. 

P 2 

Calcaires et marnes avec la faune de 
Zechstein, gypse. 

Les gres cuivreux avec les plantes 
fossiles {Chelepteris, Sphallopteris, 
Bathypteris, et Anomorrhma). 







Marnes bigarres, argiles, gres ou 
avec la faune marine : Produdus 
Cancrini, Strophalosia horrescens, 
ou avec les plantes fossiles 
( Catamites gigas). 




















Artinskien. Koungourien. 

Calcaires dolomitiques (Produdus 
Cancrini, Marginifera typica, 
Dielasma elongata, Fusulina 

Gres d'Artinsk, avec les plantes 
fossiles et ammonitides. 

* Compiled from Lapparent and other sources. 


Zalesskya gracilis, Eichwald, sp. (PI. I. figs. 1-3 ; PL II. figs. 4, 5, 8 ; 

PL III. figs. 9-20.) 

1860. Chelepteris gracilis, Eichwald, Leth&a Rossica, vol. i., p. 98., pi. iii. figs. 4-5 (1 non fig. 6). 
1869. Chelepteris gracilis, Schimper, Traite d. paleont. veget., vol. i., p. 702, pi. H. fig. 4 (1 non 
fig. 5). 

The description of this species is based upon a portion cut from the lower end of 
the type specimen figured by Eichwald in his Lethsea Rossica, Taf. iii., fig. 4. A 
photograph of the complete fossil by Mons. R. Koch was kindly forwarded to us by 
Mods. Zalkssky, and is reproduced in Plate I., fig. 1. 

As seen in transverse section, the greater part of the fossil is in an excellent state 
of preservation, although the central tissues have unfortunately disappeared. The 
most conspicuous part of the stem is the very wide stele (fig. 2, st.), which is 13 mm. 
in diameter. The stele is surrounded by a very wide thin-walled cortex, more of 
which is preserved on one side than on the other. At its widest the cortex is about 
20 mm. thick, but even there it certainly does not represent the whole thickness of 
the stem. So far as it is present, the cortex is all thin -walled parenchyma except a 
narrow border at the very periphery of its widest part (fig. 2, sc. C): At this region 
the cortical elements somewhat suddenly become thick-walled and brown-coloured. In 
all probability this represents the inner limit of an outer sclerotic cortical zone such as 
occurs in the other Osmundacese, both fossil and living. The inordinate width of the 
thin-walled inner cortex (fig. 2, i.e.) is a striking feature in our fossil. 

The cortex is traversed on all sides by a very large number of departing leaf-traces, 
which arise from the stele in a close spiral. Here and there roots are also cut across 
in various directions as they pass outwards. Each leaf-trace is accompanied by a 
sheath of thin-walled inner cortex as it passes through the sclerotic zone (fig. 2, A). 
Unfortunately, our sections do not include the extreme periphery of the stem, and there 
is no indication of the leaf-bases becoming free from one another. 

The Structure of the Xylem. 

The xylem of the stele forms a broad and perfectly continuous ring surrounding a 
central empty space (fig. 2). At one point in the stele a portion of it has been torn 
away, but elsewhere the ring measures l - 8 to 1"3 mm. in thickness. It was un- 
doubtedly still wider in the living condition, for its inner margin is bordered at many 
places by the remains of crushed and flattened elements (figs. 3 and 4, x). Some six 
or seven series of these crushed elements are present, and if they were imagined to have 
retained their original form and size, the xylem ring would be about *9 mm. wider. 
Therefore the maximum width of xylem for which we have actual evidence is about 
3 "2 mm. But at the same time the complete ring is 12 mm. in diameter, and therefore 
there remains a space 5*6 mm. wide in the middle of the stele still unaccounted for. 


There is no reason to believe that this actually represents an empty space in the living 
plant ; on the other hand, it was doubtless occupied by some kind of tissue. 

The xylem is composed of tracheides alone, without any trace of xylem parenchyma. 
The peripheral elements of the ring (figs. 3 and 4, o. xy.) are most distinctly smaller 
than the more central (figs.* 3 and 4, i. xy.), and their average size increases gradually 
towards within until a certain point is reached not far from the inner margin of the 
ring. At this point a distinct change is noticeable in the general aspect of the xylem, 
owing to the presence of a number of conspicuously larger tracheides. In fact, two 
definite zones may be distinguished in the xylem ring, although the difference between 
the two kinds of xylem is not very pronounced in transverse section. In longitudinal 
section, however, the outer xylem zone is seen to consist of very elongated elements 
with slowly tapering and often very finely pointed ends (fig. 4. o. xy.), which appear to 
have undergone much sliding growth. On the other hand, the elements of the inner zone, 
and especially the larger ones, are comparatively short, and they are somewhat contorted in 
form (fig. 4, i. xy.). They are also more or less square-ended, with transverse or at most 
slightly oblique terminal walls. Wherever the flattened elements at the inner margin are 
sufficiently well preserved, they show the same characteristics as those of the inner xylem. 

Close within the external periphery of the xylem ring occur some twenty to twenty- 
five well-defined mesarch groups of protoxylem, evenly distributed all round the stele 
(fig. 3, prx., and fig. 11). These protoxylems are really the downward prolongations of 
the protoxylems of the leaf- traces, which are decurrent for some distance into the xylem 
of the stem, gradually dying out below. It is possible to regard the small elements 
that line the outer periphery of the stele as a continuous exarch protoxylem proper to the 
stem itself, and apart from the decurrent mesarch protoxylems of the leaf- trace. At the 
same time, however, they are rather large for protoxylem elements, and they do not differ 
in form or in the nature of their pitting from the contiguous elements of the metaxylem. 

The mesarch protoxylems decurrent from the leaf-traces consist of a few very small 
tracheides, which are typically scalariform ; that is to say, each of their walls has only 
one series of pits. On the other hand, the elements of the rest of the xylem have two 
or more vertical series of pits on almost every wall. It is only on the narrower walls ; 
up to about 40 m broad, that a single series is found (fig. 5 a). A more frequent 
marking is that of two regular vertical series of more or less horizontally elongated 
oblong pits (figs. 5 b and c). The unpitted region separating the two series varies 
considerably in breadth, and should it exceed a certain limit, a third series of pits 
appears in. the middle of the wall (fig. 5 d). The pits of this middle series are 
usually smaller and more irregular than those of the lateral, being oval or almost 
rounded in outline. More rarely three or even four regular series of pits, all similar in 
size and form, are met with (figs. 5 e and/"). When the number of series is still greater 
the pits are usually quite irregular in arrangement and oval or elliptic in outline 
(figs. 5 g and h). In the large elements of the inner zone of xylem this irregularity in 
arrangement is so great and the pits are so numerous that the thickened part of the 



wall separating them appears as a delicate reticulation. In fact, in this last case they 
might very well be described as porose. The end walls of the tracheae are marked with 
the same pattern as the side walls. In the transverse section of this fossil the tracheal 
walls are represented in such a manner that a considerable amount of structural detail is 
visible in the substance of the wall itself. In its best state of preservation the tracheal 
wall is represented by a brown-coloured substance showing up in the matrix of the 
fossil. The brown tract is delimited from the cavities of the tracheides that it separates 
by very fine black lines. These are usually most conspicuous at the angles and at the 
unpitted regions of the wall (figs. 8 b, c, and d), but they are often more or less broken 
up. and may even be absent altogether. A varying number of thin black lines also occur 
in the median region of the brown tract that represents the wall (figs. 8 a-d). These 
are of different lengths and are separated from one another by the homogeneous brown 
substance of the tract in which they lie. Similar black lines also occur at the angles 
where three or more tracheae meet (figs. 8 a-d). They may either meet at the very 
centre of the angle to form a three- or four-rayed figure, or they may delimit one or more 
sides of an angular mass of the homogeneous brown substance occupying that position. 
In either case they radiate out shortly into the walls that meet at the angle, but 
they never become continuous with the other dark lines that lie in the substance 
of these walls. 

From comparison with longitudinal sections it is clearly seen that the longer of 
these black lines in the median region of the wall correspond to the several vertical 
series of pits. Thus, if there is one series of pits, there is a single long black line 
(fig. 8 a) ; if two or more series of pits, there are two or more black lines of equivalent 
length (fig. 8 b). Here, however, a difficulty arises, for very often a very short line is 
also present between the longer lines which correspond to the pit series, and this can- 
not be referred to any pit at all (figs. 8 c and d). In many parts of the fossil the walls 
sometimes appear to have become more or less disorganised and dissolved before 
fossilisation. This breaking down usually occurs in the middle of the wall, and affects 
either an unpitted region or else the middle series of pits (fig. 8/). Sometimes the 
wall is broken down at several places, or even along its whole length (fig. 8 e). In these 
regions the brown-coloured substance is partially or entirely absent, and the wall is 
represented by a thin black line alone. 

The peculiarities just described in Zalesskya gracilis and also others noted in the 
species of Osmundites dealt with in Part I. of this paper were so difficult to correlate 
with the generally accepted idea of the structure of a fern tracheide that a careful 
re-examination of the xylem of the living Osmundacese was rendered necessary. This 
investigation was carried out by one of us, and the results, which are here shortly 
summarised, will be published in detail as a separate paper in a forthcoming number 
of the Annals of Botany. Contrary to expectation, it was found that the middle 
substance of the mature wall separating two contiguous tracheae is entirely wanting 
in the regions of the pitted areas. In fact, the middle lamella and the primary walls 



of the young tracheae only persist at the angles where the tracheae meet and in the 
areas separating the vertical series of pits (fig. 6). The opposing bars of secondary 
thickening that separate the several pits of each vertical series are quite free from one 
another in the median vertical plane in the long axis of the wall. If there is only 
one series of pits, the opposing bars of the two contiguous tracheae run freely across 
from one corner to the other without coming into contact with one another. If there 
are several vertical series of pits, the bars are cemented together by an intervening 
middle substance at the intervals between each series of pits. 

In the transverse section of such a wall an empty space will be seen in the middle 
of its substance corresponding with the position of each vertical series of pits, whereas 
the angles and the intervals between the series of pits will be solid throughout (fig. 6). 
In applying these facts to the fossil Osmundacew, no difficulty is met with in the cases 
of Osmundites Dunlopi, 0. Gibbiana, and 0. skidegatensis. In these the walls are 
represented by their dark carbonaceous remains, and wherever detail is preserved 
their appearance fits in accurately with the description of the recent Osmundacese 
given above. This is best shown in Osmundites skidegatensis (fig. 7), where the 
transverse section of the tracheal wall shows two black bars separated by white spaces 
in the pitted regions, but connected up by a black mass in the unpitted regions and 
at the angles where the trachea? meet. 

We feel satisfied that in the living Zalesskya gracilis the tracheal wall also possessed 
the same structure, although it is somewhat masked in the fossil owing to some peculiarity 
in its method of fossilisation. Indeed, in some of the sections a structure similar to that 
of Osmundites skidegatensis is clearly present in the root-steles and leaf-traces (fig. 20). 
Elsewhere it appears that the disintegrated carbonaceous substance of the wall, during 
the processes of partial decay and fossilisation, has become redeposited on the surfaces 
of the wall, both on the internal surfaces of the lumens of the tracheae and on the 
limiting surfaces of the empty spaces in the substance of the walls, producing the thin 
black lines referred to above. This still leaves the black lines at the angles and the 
short lines opposite unpitted areas in the wall unexplained. It does not seem probable 
that these represent cavities existing in the living material, but they may indicate the 
position of tracts of some less resistant substance than lignin in the tracheal wall, 
which, by its early decay after the death of the plant, formed cavities in the wall 
before actual fossilisation took place. 

Structure of Phloem and Cortex. 

The xylem is surrounded by a continuous ring of phloem (figs. 3 and 9, ph.), separ- 
ated from the tracheae by a broad zone of some five to six layers of vertically elongated 
parenchymatous cells forming the xylem sheath (figs. 4 and 9, xy. sh.). Some of these 
cells are filled with a dark brown substance which sometimes appears granular, and the 
innermost often have their angles marked by dark lines similar to those that occur at 
the angles of the tracheae. Probably the walls at these points were more or less 


lignified, as is actually the case in the corresponding cells of Osmunda cinnamomea. 
The phloem ring is a broad one, and consists for the most part of large and conspicuous 
sieve-tubes (figs. 9 and 10), amongst which a few elements are interspersed which may- 
be regarded as phloem parenchyma (fig. 10, ph. par.). These are much smaller than 
the sieve-tubes in size, and have somewhat firmer walls. The sieve-tubes were very 
elongated and tubular (fig. 4, ph.), with stout but somewhat soft walls, for in most places 
they are much distorted and often pressed flat, so that the lumen is almost obliterated. 
Where they are well preserved, a thin brown outline is visible inside the sieve-tube, 
suggesting a pellicle of some substance that has separated off from the wall and become 
more or less contracted and folded (fig. 10, pel.). It may have been the inner layer of 
the wall, or possibly represents the contents. No suggestion of a " porose layer" or a 
protophloem could be detected in any part of our sections, transverse or longitudinal. 
On the contrary, towards the outside, the phloem is directly continuous with a zone 
of four or five layers of parenchymatous cells, conspicuous in the fossil owing to the 
dense dark brown colour of their contents (figs. 4 and 9, per.). These contents are not 
homogeneous, but exhibit a curious vesicular structure varying in texture from a 
coarse foam to one so fine that it might almost be described as granular. The cells of 
this layer are angular in outline, and fit into one another without leaving intercellular 
spaces. Those of the innermost layers are comparatively small and vertically elongated ; 
towards without they increase in size, becoming at the same time shorter, and thus 
they pass over gradually into the parenchyma of the inner cortex. This innermost 
zone must be regarded as a pericycle, although it is admittedly not marked off from the 
cortex by any definite single layer comparable to an endodermis. At the same time, 
the stele as a whole has a fairly sharp external contour, for the cells of the inner cortex 
distinguish themselves from those of the pericycle by the general absence of the dense 
brown contents and by the presence of well-marked intercellular spaces at the corners 
(fig. 3). The broad inner cortex consists of thin- walled parenchymatous ovoid or rounded 
cells, with intercellular spaces at the angles. In most of the cells the contents have been 
completely disintegrated, but a considerable number are filled with a brown vesicular 
substance similar to that described in the cells of the pericycle. Judging from the 
contents of the cells in a corresponding position in the recent Osmundacese, and also in 
some of the fossil forms (Osmundites skidegatensis, 0. Doivkeri), the vesicular masses 
probably represent aggregates of starch grains in various grades of decomposition. 
This is the more probable because in those parts of the cortex where the cells are best pre- 
served the vesicular mass resolves itself into a number of separate contiguous spherical 

At the periphery of the widest part of the cortex the cells gradually become 
narrower and somewhat more vertically elongated. At the same time their walls 
become much thicker and dark brown in colour (fig. 2, Sc. C). This region doubtless 
represents the inner margin of a sclerotic outer cortex, the greater part of which is 
wanting in our specimen. 


The Leaf-trace. 

The leaf-trace departs from the stele of the stem in a perfectly protostelic manner. 
The protoxylem groups previously described (fig. 11) at the periphery of the xylem 
ring are all decurrent from the leaf-traces. The first sign of preparation for the de- 
parture of the leaf-trace is the appearance of a slight prominence on the periphery of 
the xylem ring opposite one of the mesarch protoxylems (figs. 12 and 13, prx.). Followed 
upwards, this develops into a broad, bluntly rectangular projection into which the 
protoxylem group passes (fig. 13). This projection gradually increases in size and 
eventually separates off as the xylem strand of a leaf-trace (fig. 14). While still enclosed 
within the phloem of the stem, the protoxylem of the leaf- trace is seen to be deeply 
mesarch, with a large amount of centripetal xylem on its adaxial side (fig. 14). Even 
after the leaf-trace as a whole has become completely free from the stele and has advanced 
some distance into the cortex, a considerable number of centripetal elements are still 
present (fig. 15, a and b). 

In the close neighbourhood of the stele the transverse section of the xylem of the 
leaf- trace is oblong in outline, while the leaf- trace itself is oval (fig. 16). As it passes 
outwards both the xylem strand and the whole leaf-trace become curved round aclaxially. 
This curvature gradually increases until in the outermost trace in our specimens the 
xylem strand has attained the form of a semicircle, while the leaf-trace itself is reniform 
(fig. 17). As the leaf- trace passes out the centripetal xylem gradually disappears, and 
by the time that the curve of the xylem strand has become at all pronounced it has 
entirely gone and the protoxylem is perfectly endarch (fig. 16, prx.). In a leaf-trace 
still further out the protoxylem group is seen to have divided, and in the outermost 
leaf- traces two or three protoxylem strands are present. No doubt the leaf-trace became 
still more curved and the protoxylems still more numerous in the outer region of the 
sclerotic cortex that is, however, wanting in our specimen. 

The xylem of the leaf-trace is surrounded on all sides by a parenchymatous sheath 
two to four layers thick, and similar in appearance to the xylem sheath of the stem 
(fig. 16. xy. sh.). This tissue is followed by a continuous and well-developed zone of 
phloem consisting of large sieve-tubes and a few cells of phloem parenchyma. On the 
other hand, the closest examination failed to reveal any protophloem whatever, even on 
the abaxial side of the trace. While still in close proximity to the stele of the stem 
the leaf-trace is limited by a zone of some three or four layers of cells exactly similar 
in appearance to those that have been referred to as the pericycle in the stem. As the 
leaf-trace passes outwards, the contents of the inmost cells of this zone become less and 
less conspicuous, while those of the outermost layer become more dense and prominent. 
In fact, this layer gradually assumes the appearance of an endodermis (fig. 16, en.). 

The Root. 
The roots arise singly or in pairs upon the stem stele at points immediately below 
the departure of certain of the leaf-traces. Many traces, however, are not provided 


with roots. The xylem strands of the roots are inserted upon the sides of those of the 
leaf-traces just as the latter separate from the xylem of the stem (fig. 18, r. st.). The 
roots run more or less horizontally through the cortex, taking a sinuous course and 
turning aside from time to time to avoid a leaf-trace or another root. 

When about half-way through the inner cortex of the stem they obtain a cortex of 
their own, which consists for the most part of very thick-walled fibrous sclerenchyma 
(fig. 19, se. v.). The cells are smallest and thickest at the outside ; towards within they 
increase in size, while their walls diminish in thickness. The last one or two layers next the 
endodermis are comparatively thin-walled. The cells of the endodermis are tangentially 
elongated and filled with brown, densely granular contents (fig. 20, tn.). It probably 
consists of a single layer of cells, but it is difficult to make sure of this, for the adjacent 
cells sometimes contain an exactly similar substance. 

The xylem strand is diarch and oval, with a few very small protoxylem elements at 
the two ends. The xylem is surrounded by a parenchymatous sheath two or three cells 
thick, and this is followed by the phloem, which is in considerable quantity on the sides 
of the xylem. It is possible also to distinguish protophloem on the outside of the two 
phloem groups (fig. 20, pr. ph.). The phloem is separated from the endodermis by two 
or three layers of pericycle similar to that of the stem in general appearance. 

Locality. — Mine Kluczersky, district of Bjelebei, government of Orenburg, Russia. 

Horizon. — Upper Permian (P2 of Geological Table, p. 219). Specimen preserved 
in the Museum of the Institute of Mines, St Petersburg. 

Zalesskya diploxylon, Kidston and Grwynne-Vaughan, n.sp. 

(PL IV. figs. 21-28.) 

The general appearance of this fossil suggests that it has undergone a considerable 
amount of change after mineralisation, owing to the percolation of water through the 
stone. The sections were of a light fawn colour, much of the carbonaceous matter 
originally present having been removed. However, a certain amount of this has been 
redeposited in the form of minute granules, flakes, and specks, and in some parts 
of the fossil as large irregular aggregations. The substance of the fossil as it exists at 
present is very porose and friable. 

The transverse section of the specimen is circular, with a diameter of about 3*5 cm. 
(fig. 22). There is a fairly large stele measuring 7 mm. across. The greater part of the 
xylem is still preserved, but unfortunately the more central elements of the stele have 
crumbled away, owing to the friable nature of the fossil. The xylem is surrounded by 
a rather conspicuous ring of phloem, visible to the naked eye, and this in turn by a 
very wide cortex entirely composed of thin-walled parenchyma. In our specimen the 
whole of the outer part of the stem, and also the mantle of persistent leaf-bases that 
probably coated the stem proper in the living plant, have; not been preserved. It 
follows, therefore, that the surface of the fossil as shown in fig. 21 represents a 


" Knorria" or partially decorticated condition of the actual stem. The true external 
surface of the stem, and still more so that of the coating of leaf-bases, would present a 
very different appearance. Numerous leaf- traces and root-steles are cut through on 
their way out through the cortex. The former arise in a close spiral upon the stele. 

Structure of the Stele and Cortex. 

As in Zalesskya gracilis, the xylem forms a perfectly continuous ring and the leaf- 
traces depart in a protostelic manner. The break in the xylem seen to one side of the 
stele (fig. 22) is obviously due to the crumbling away of the stone. This also accounts 
for the absence of tissue in the centre of the stele. Fig. 23 shows clearly that the 
broken inner margin is not the natural limit of the xylem, but the same tissue extended 
further in towards the centre of the stele. The whole mass of the xylem is 6 '4 mm. 
in diameter, but the widest part of it still preserved only measures 2*5 mm. across. If, 
however, an equal amount of it had been preserved all round, the central area still 
unaccounted for would only be 1 '4 mm. wide. 

As in Zalesskf/a gracilis, xylem parenchyma is entirely absent, and two distinctly 
different regions can be observed in the xylem — a central and a peripheral (figs. 23 and 
24). The contrast between the two kinds of xylem appears to be much more marked in 
Zalesskya diploxylon than in Z. gracilis, but this is perhaps due to the fact that we 
have so much more of the central xylem present in Z. diploxylon than in the other 
species. The elements of the outer xylem zone are very elongated, with more or less 
pointed ends. Those at its extreme periphery are not much smaller than the rest, but 
just where the outer zone borders on the central xylem there is a somewhat sudden and 
fairly distinct decrease in the general size of the elements (fig. 23, at x). The elements 
of the central xylem are much wider than those of the outer zone, and in longitudinal 
section the contrast between the two kinds of xylem is even more strongly marked than 
in the transverse (fig. 24). This is chiefly due to the fact that, in addition to being- 
wider, the elements of the central xylem are also much shorter than those of the outer 
zone, and further, their terminal walls are transverse, or at most only slightly oblique 
(figs. 24 and 25). In fact, the tracheae of the central xylem might well be described as 
inflated, oblong, or sack- shaped ; some of the inner ones not being more than twice as 
long as broad. 

The elements of both kinds of xylem have multiseriate pits, those of the outer 
zone having usually two vertical series of oblong pits ; but in the central xylem as many 
as five regular series were observed. Where the pits are more numerous than this they 
are more or less irregularly arranged and the trachese have a reticulate or porose 
appearance. In spite of their parenchymatous and irregular form there is no doubt that 
even the inmost of the elements preserved belonged to the xylem, being in fact true 
pitted tracheae. Further, we are strongly of the opinion that in the living plant similar 
elements filled up the whole of the vacant space that now occurs in the centre of the 
stele. In fact, we believe that the xylem was solid throughout. 


As usual, the xylem is surrounded by a broad sheath of parenchyma some four to 
five layers thick (figs. 23 and 26, xy. sh.). Then follows a stout belt of phloem, con- 
sisting of from four to six layers of sieve-tubes without any intervening phloem 
parenchyma (fig. 26, ph.). The sieve - tubes are large and very elongated, and 
probably had fairly thick and firm walls, for in most parts of the section they have 
kept their form without much distortion. As in Zalesskya gracilis, nowhere in any 
of our sections is there any indication either of a protophloem or of any " porose 
layers." On the contrary, the large sieve-tubes abut directly upon a zone of 
parenchymatous cells occupying the position of a pericycle (fig. 26, per.). The latter 
is marked off from the cortex by an almost continuous ring of irregularly shaped 
aggregations of carbonaceous matter (fig. 26, car.), but no endodermis is recognisable. 

The cortex consists throughout of oval or rounded parenchymatous cells with inter- 
cellular spaces at the corners. No remains of an outer sclerotic cortex have been left 
at any part of its periphery. 

The Leaf-trace. 

The first sign of the departure of a leaf- trace from the stele is the appearance of a 
slight prominence on the surface of the xylem. This projection increases more and 
more towards above (fig. 27), until at last it is constricted off as the oblong xylem 
strand of the trace. In a leaf-trace some little distance from the stele the xylem 
strand is slightly reniform in outline, with a single median endarch adaxial protoxylem 
(fig. 28, prx.). The leaf-trace protoxylem is decurrent for some distance into the 
xylem of the stem as a mesarch strand, but it dies out much more rapidly than in 
Zalesskya gracilis. In fact, it has already disappeared by the time that the prominence 
of the leaf-trace xylem is no longer recognisable. In our sections the leaf-traces in 
closest proximity to the stele have all fallen out, so it could not be determined whether 
any centripetal xylem accompanied the leaf- trace from the stem stele or not. 

As the leaf-trace passes outwards the curvature of its xylem strand gradually in- 
creases until in the outermost traces it has assumed the form of a low arch. At the same 
time the median protoxylem divides first of all into two and ultimately into several 
separate strands. The outline of the whole leaf-trace passes through a similar series of 
changes, although it always remains considerably less curved than its own xylem strand. 

In several of the leaf-traces the tissues are fairly well preserved, and it can be made 
out that there are two or three layers of parenchyma separating the xylem from the 
phloem (fig. 28). The phloem, which completely surrounds the leaf-trace, is in greatest 
abundance in the median region on both sides of the trace, and particularly on the 
adaxial side, where it may consist of as many as six or seven layers of sieve-tubes. As 
in the stem, it appears to consist of sieve-tubes only without any phloem parenchyma. 
The preservation was not sufficiently good to determine the presence or absence of a 
protophloem. Around the phloem are three or four layers of cells, which may be 


regarded as a pericycle, and the limit of the whole leaf-trace is marked by a ring of 
black carbonaceous matter (fig. 28, car.). 

The Boot. 

The roots arise singly or in pairs directly on the stele. Their xylem strands are 
attached directly to the xylem of the stele in a line with, but below, the point of 
departure of a leaf-trace xylem. On their way through the cortex they first of all run 
towards the apex of the stem (fig. 22, ?'.), and then they turn outwards and their 
course becomes very irregular. Their xylem strands are diarch, and they do not attain 
a cortex of their own, so far as could be observed in our specimen. 

Locality. — Found in the same collection as Zalesskya gracilis, Eichwald, sp., but 
without label. The specimen is, however, believed to have come from the same 

Specimen preserved in the Museum of the Institute of Mines, St Petersburg. 

Zalesskya, n.g. 

Fern stems of considerable size. Leaves spirally arranged. Axis occupied by a 
single central protostele. Xylem forming a stout continuous ring, or solid to the 
centre. Xylem differentiated into two regions — a peripheral zone of normal tracheae 
and a more central region of short, wide vesicular elements. Tracheae with multiseriate 
pits. Phloem well developed, separated from the xylem by a stout xylem sheath. 
Protophloem and porose layers absent. Stele surrounded by a very wide parenchy- 
matous cortex ; outer cortex probably sclerotic. Leaf- trace and its xylem strand at 
first ovoid in transverse section, with a median mesarch protoxylem near the adaxial 
side ; further out horse-shoe shaped, with the concavity adaxial and with several 
endarch protoxylems. 

General Considerations. 

The most interesting points brought to light by the investigation of the genus 
Zalesskya are, first of all, the absolute confirmation of the occurrence of a completely 
continuous and perfectly protostelic xylem ring in the vascular anatomy of the 
Osmundacese, and, secondly, the marked distinction that exists between the elements 
of the peripheral and those of the more central regions of the xylem. The peripheral 
xylem zone is obviously represented in the recent Osmundacese by the ring of xylem 
strands that occurs in the stele, but the central xylem, with its short, wide reticulate or 
porose elements, appears to have been lost in the existing forms. 

As regards Zalesskya diploxylon, at any rate, we feel convinced that the central 
xylem occupied the whole of the centre of the stele in the living plant. Further, we 
accept the deduction suggested by this conclusion, that the vascular anatomy of the 
Osmundacese must he derived from a protostele with a solid central homogeneous 



xylem mass. In fact, these fossils seem to disclose some of the steps by which the 
transition actually took place. In Zalesskya diploxyJon the central trachea? have 
become short and wide, and seem to have changed their water-carrying function for one 
of water-storing. In the recent and the more advanced forms they have become still 
further modified, and, losing all trace of their original tracheal nature, they have 
given rise to a thin- walled parenchymatous pith. From this point of view it is 
clear that the central ground- tissue of the recent Osmundacese must be regarded as 
phylogenetically derived by modification from the central xylem of a solid protostele, 
and that primitively it had no relation or connection ivith the cortex ivhatever. The 
vascular anatomy of the Lepidodendrese and Sigillarise provides an instructive and 
closely parallel series of developments. A similar advance is made, starting from a 
homogeneous solid mass of xylem (L. rhodumnense, etc.), through such a form as 
L. vasculare, which has a pith intermingled with short isodiametric tracheae, leading on 
to forms with a pure pith (L. Harcourtii, etc.). The next stage is the partial 
breaking up of the narrow xylem ring as seen in Sigillaria spmidosa (Scott, Studies 
in Fossil Botany, p. 200), until finally a stage is reached in which the xylem strands 
are completely separate from one another, as in Sigillaria Menardi (Brongniart, 
Observations sur la Structure du Sigillaria elegans, pi. xxv., figs. 3 and 4 ; pi. xxvii., 
fig. 1). At this point reference should be made to an interesting observation by 
Seward and Ford in their paper on the anatomy of Todea (Trans. Linn. Soc. Lond., 
vol. vi., pt. 5, 1903, pp. 248 and 249). They record the occasional occurrence of short, 
wide tracheae with reticulate thickenings at the inner margins of the xylem strands of 
Todea superba and T. hymenophylloides. The former is figured on pi. xxix., figs. 
30 and 31. These elements may perhaps be regarded as the vestigial remains of the 
ancestral central xylem. 

The peripheral tissues of the stele in the living Osmundacese possess some special 
characters of exceptional interest and peculiar to the order — in particular, the absence 
of a true protophloem and the presence of the so-called " porose layers " on the outside 
of the metaphloem. In Zalesskya both these tissues are absent, and the large sieve- 
tubes of the metaphloem abut directly upon a zone of parenchyma that we have called 
the pericycle. It appears, therefore, that the porose layers are a relatively late develop- 
ment, and that Seward and Ford are right (I.e., p. 242) in regarding it as derived from 
the pericycle. In the Osmundaceous affinity the phloem of the stem has never become 
differentiated into protophloem and metaphloem, not even in the most advanced forms. 
On the other hand, a well-developed protophloem is present in the leaf-trace of the 
recent genera and also in Osmundites skidegatensis. 

As regards the zone of tissue that we have called the pericycle in these two fossils, 
it only deserves the name in virtue of its position. No real delimitation is possible 
between its cells and those of the inner cortex, and there is no definite layer that can 
be identified as an endodermis. In fact, the state of affairs in these primitive forms would 
be best expressed by saying that the phloem of the stele was surrounded by a narrow 


sheath formed by slightly specialised cells of the inner cortex. An endodermis might 
have been demonstrable in the living plant as in the modern genera, and whenever it is 
present it must have been differentiated from the outer cells of this sheath, which may 
account for the slight irregularities which it frequently shows in the recent forms. These 
observations on the phloem and pericycle are, of course, far too limited to do more than 
provide mere suggestions as to the origin of these tissues. A complete knowledge of 
this region of the stele in other primitive members of the order is necessary before 
any confident conclusions can be reached. 

Owing to the decorticated nature of our specimens, it is unfortunately impossible to 
say whether the base of the petiole possessed stipular wings or not, and no comparisons 
can be made with the structure of the petiole in the other members of the Osmundacese. 
It should be mentioned, however, that in Zalesshya there is no alternation of scale- 
leaves and foliage-leaves such as occurs in the recent genera and in the Jurassic 
representatives Osmundites Dunlopi and 0. Gibbiana. The mesarch position of the 
protoxylem in the leaf-trace when still in close proximity to the stem stele is, we believe, 
a point of considerable importance ; but since another Osmundaceous fossil to be described 
in the next part of this paper promises to throw some light upon the point, it will be 
best to leave it over for future consideration.* 


Figs. 8 a-f are from drawings ; all the others are from untouched photographs. 

The following lettering is used throughout: xy., xylem; prx., protoxylem; ph., phloem; S.T., sieve- 
tubes; per., pericycle; C, cortex; l.t., leaf-trace; r., root. 

Plate I. 

Fig. 1. Zalesshya gracilis. External surface of the whole fossil. Nat. size. (Photograph by Mons. R. 
Koch; forwarded by Mons. M. Zalessky.) 

Fig. 2. Znlesskya gracilis. A complete transverse section, st., stele; Sc. C, sclerotic cortex; i. C, 
parenchymatous cortex ; A., tract of inner cortex accompanying the leaf-trace through the sclerotic outer 
cortex. x about 2J. (Slide 1253.) 

Fig. 3. Zalesshya gracilis. Portion of the vascular ring in transverse section, o. xy., outer xylem : 
i. xy., inner xylem. The black band at x represents crushed tracheides. x 30. (Slide 1253.) 

Plate II. 

Fig. 4. Zalesshya gracilis. The vascular ring in longitudinal section, xy. sh., xylem sheath. x 30. 
(Slide 1258.) 

Fig. 5. Zalesshya gracilis. Surface views of the tracheal walls as seen in longitudinal section, a, a 
single vertical series of pits ; b, two such series ; c, two series widely separated ; d, three series, the median 
one being smaller than the lateral ; e, three series all similar ; /, four equal series ; g, reticulate or porose 
pitting; h, terminal wall between two reticulate or porose elements. x 250. (Slides Nos. 1255, 1256, 
1257, 1258.) 

* We beg to acknowledge our indebtedness to the Executive Committee of the Carnegie Trust for a grant to 
defray the cost of the plates illustrating Parts I. and II. of this paper. 


Fig. 6. Oismunda cinnamomea. Wall between two xylem elements. x 600. 

Fig. 7. Osmundites skidegatensis. Transverse section of the xylem showing the structure of the walls of 
the tracheae, x 180. (Slide 1251.) 

Fig. 8. Zalesskya gracilis. Transverse sections of the walls of the xylem elements, a, a single series 
of pits, each represented by a black line in the substance of the wall ; b, two series of pits ; c, ditto with an 
additional small black line; d, three series of pits with additional small black lines; e and/, imperfectly 
preserved walls. x about 600. 

Plate III. 

Fig. 9. Zalesskya gracilis. Transverse section of the peripheral tissues of the stele, xy. sh., zone of 
parenchyma separating xylem from phloem; in. c, inner cortex. x 50. (Slide 1252.) 

Fig. 10. Zalesskya gracilis. Transverse section of the phloem; S. T., sieve-tubes; pel., pellicle within 
the sieve-tubes; ph. par., phloem parenchyma; xy. sh., xylem sheath. x 260. (Slide 1252.) 

Fig. 11. Zalesskya gracilis. Transverse section of the periphery of the xylem showing a mesarch 
protoxylem group. x 80. (Slide 1253.) 

Fig. 12. Zalesskya gracilis. Transverse section of the periphery of the xylem showing the first 
indication of the departure of a leaf-trace. x 30. (Slide 1253.) 

Fig. 13. Zalesskya gracilis. The departing leaf-trace at a point higher up than that shown in fig. 12. 
x 30. (Slide 1235.) 

Fig. 14. Zalesskya gracilis. The leaf-trace with its xylem just free from the xylem of the stele. x 30. 
(Slide 1252.) 

Fig. 15 a and b. Zalesskya gracilis. Transverse sections of two different leaf-traces in close proximity 
to the stele. The protoxylem is still mesarch. x 40. (Slide 1253.) 

Fig. 16. Zalesskya gracilis. Transverse section of a leaf-trace some distance out in the cortex, xy. sh., 
xylem sheath ; ad. ph., adaxial phloem ; ah. ph., abaxial phloem ; era., endodermis. x 90. (Slide 1253.) 

Fig. 17. Zalesskya gracilis. Transverse section of one of the outermost leaf-traces in the specimen, 
x about 12. (Siide 1253.) 

Fig. 18. Zalesskya gracilis. Transverse section of the periphery of the stele showing the departure of 
the root steles; r. st., root steles. x 20. (Slide 1252.) 

Fig. 19. Zalesskya gracilis. Transverse section of a root when still in the cortex of the stem; sc. r., 
sclerotic cortex of root, x 22. (Slide 1260.) 

Fig. 20. Zalesskya gracilis. Transverse section of the stele of the same root more highly magnified. 
pr. ph., pro to phloem ; en., endodermis. x 110. (Slide 1260.) 

Plate IV. 

Fig. 21. Zalesskya diploxylon. External surface of the specimen. Nat. size. (Photograph by Mons. 
R. Koch; forwarded by Mons. M. Zalessky.) 

Fig. 22. Zalesskya diploxylon. A complete transverse section of the specimen, st., stele, x about 2£. 
(Slide 1268.) 

Fig. 23. Zalesskya diploxylon. Transverse section of a portion of the stele, o. xy., outer xylem ; 
cen. xy., central xylem ; x., inner limit of outer xylem zone. x 30. (Slide 1268.) 

Fig. 24. Zalesskya diploxylon. Longitudinal section of the stele. Lettering as in fig. 23. x about 14. 
(Slide 1271.) 

Fig. 25. Zalesskya diploxylon. Longitudinal section of a portion of the central xylem. x 35. 
(Slide 1271.) 

Fig. 26. Zalesskya diploxylon. Transverse section of the peripheral tissues of the stele, car., zone of 
carbonaceous matter, x 65. (Slide 1268.) 

Fig. 27. Zalesskya diploxylon. Transverse section of the periphery of the stele showing a departing 
leaf-trace. x 30. (Slide 1268.) 

Fig. 28. Zalesskya diploxylon. Transverse section of a leaf-trace in the cortex of the stem. ad. ph., 
adaxial phloem ; ah. ph., abaxial phloem ; car., ring of carbonaceous matter. x 30. (Slide 1269.) 

Trans. Roy. Soc. Edinr Vol. XLVI. 


Sc. C. 

Zalesskya gracilis. Eichwald sp. 

rans. Roy. Soc. Edinr Vol. XLVI. 


Figs. 4, 5. Zalesskya gracilis. Eichwald sp. Fig. 6. Osmunda cinnamomea. Linn. 

Fig- 7. Osmundites skidegatensis. Penhallow. Fig. 8. Zalesskya gracilis. Eichwald sp. 

Trans. Roy. Soc. Edinr 

Vol. XLVI. 


Zalesskya gracilis. Eichwald sp 

Trans. Roy. Soc. Edinr Vol. XLVI. 

Kidston and Gwynne-Vaughan : Fossil Osmundace^:. Part II. Plate IV. 





w tw i *C«TE V 


xy. sh 

* "'i ' " 

r.r. ■ "'■ ' * t ' f<i 

4 « * 


f Ct ^27. 



o.xy. 24.. cen.xy. 

Zalesskya diploxylon. Kidst. and G-V. 

( 233 ) 

X. — The Medusas of the Scottish National Antarctic Expedition. By Edward 
T. Browne, Zoological Eesearch Laboratory, University College, London. 
Communicated by Dr W. S. Bruce. (With Two Plates.) 

(MS. received November 27, 1907. Read December 16, 1907. Issued separately May 1, 1908.) 

This is a report upon the Medusse collected by the Scottish National Antarctic 
Expedition during the voyage of the Scotia in the years 1902, 1903, and 1904, 
under the leadership of Dr W. S. Bruce, to whom I am indebted for the opportunity 
of examining the specimens. 

The collection is a small one containing twelve species, but five more species would 
have probably been added if only the specimens had been in a recognisable condition. 
Some of the specimens are in excellent condition, and these were mostly taken at the 
surface, but others are very much damaged. Unfortunately, the damaged specimens 
are just the ones wanted in good condition, as they mostly come from the Antarctic 
region and from the stations where the nets were sent down to a great depth. The 
damage is done in the nets, and other collections which have passed through my hands 
tell the same tale. It is the rubbing together of the sides of the net that tears the 
medusae to pieces. All deep-sea nets should be so constructed that the sides of the nets 
cannot come together, and also provided with a large can at the end. 

The following is a classified list of the species taken on this expedition : — 



Hippocrene rnacloviana (Lesson). Falkland Islands. 
Willia mutabilis, Browne. Falkland Islands. 


Staurophora falklandica, n. sp. Falkland Islands. 
Phialidium simplex, Browne. Falkland Islands. 


Halicreas papillosum, Vanhbffen, var. antardicum, nov. Antarctic Ocean. 
Botrynema brucei, n. g. et n. sp. Antarctic Ocean. 



Atolla chuni, Vanhbffen. South Atlantic. 
Atolla wyvillei, Haeckel. Antarctic Ocean. 

Pelagia perla (Slabber). North Atlantic. 
Desmonema chierchiana, Vanhbffen. Falkland Islands. 
Phacellophora ornata (Verrill). South Atlantic. 
Aurelia solida, Browne. North Atlantic. 



The principal interest in this collection should be in the specimens taken in the 
Antarctic Ocean. There is, however, a difference of opinion about the position of the 
northern boundary line of the Antarctic Ocean, and I consider the Falkland Islands to 
be outside the boundary line. 

It will be seen from the above list that only three species have been taken 
indisputably within the Antarctic region. Halicreas papillosum, var. antarcticum, 
may turn out to be a new species when better specimens have been examined. The 
new genus Botrynema has a well-marked character in the arrangement of the margin 
of the umbrella into lobes, and the tentacles into groups. Atolla ivyvillei had already 
been recorded for the Antarctic. These three species probably live at a considerable 
depth, and belong to the deep-sea medusae. Amongst the unrecognisable specimens 
taken within the Antarctic region are a Ber'oe, and a Pleurobrachia, or an allied 

During the stay of the Scotia in Stanley Harbour, Falkland Islands, a new species 
of Staurophora was secured, very much like the North Atlantic species, and also a fine 
specimen of Desmonema chierchiana, which I have but little doubt has been described 
under several other names, and is a common Cyaneid of the Magellan Straits and the 
Falklands. In the large bottle containing the Desmonema were some macerated 
specimens of Hippocrene macloviana, Willia mutabilis, and Phialidium simplex, all 
of which are common medusae in Stanley Harbour. 

The occurrence of Phacellophora ornata in the South Atlantic is of geographical 
interest, as it has only been previously recorded from Eastport (Maine, U.S.A.), on the 
North Atlantic coast of America. The finding of Aurelia solida at a mid-North 
Atlantic station is also of interest, because quite recently it was described as a new 
species from the Indian Ocean. 


Family Margelid^;. 
Hippocrene macloviana (Lesson), 1829. 

Hippocrene macloviana, Haeckel, 1879, p. 103, Taf. v. fig. 1-2; Browne, 1902, p. 278. 

Station. — Stanley Harbour, Falkland Islands, 7th January 1903. 

Two adult specimens in a macerated condition were found in the bottle containing 
Desmonema chierchiana. 

H. macloviana is a very common medusa in Stanley Harbour, and has not been 
recorded from any other locality. 


Family Willid^e. 

Willia mutabilis, Browne, 1902. 
Willia mutabilis, Browne, 1902, p. 280. 

Station. — Stanley Harbour, Falkland Islands, 7th January 1903. 

A single specimen in bad condition was found in the bottle containing Desmonema. 
This was one of the new species collected by Mr R. Vallentin in Stanley Harbour 
during 1898-1899. 


Family Laodicehle, L. Agassiz, 1862. 

Character of the Family. — Leptomedusse with cordyli, commonly called sensory 
clubs, on the margin of the umbrella. (Browne, 1907.) 

Stauro'phora, Brandt, 1835. 

Generic Character. — Laodiceidse with four radial canals ; with a narrow cross- 
shaped stomach, and mouth extending across the sub-umbrella ; with ocelli on the 
basal bulbs of the tentacles. (Browne, 1907.) 

Staurophora falklandica, nova species. (Plate I. figs. 1-8.) 

Station. — Stanley Harbour, Falkland Islands, 7th January 1903. 

The collection contains a single specimen, which is in fairly good condition, but 
rather damaged in places on the margin of the umbrella. 

Description. — The umbrella is very thin, and measures 90 mm. in diameter when 
completely flattened out. Its natural shape would probably be something like a 
shallow watch-glass. The velum is extremely narrow for the size of the umbrella, about 
1 mm. in width, and is in an almost rudimentary condition. 

The stomach forms a large perradial cross, which extends nearly to the margin of 
the umbrella, and consequently the true radial canals are very short, about 6 mm. in 
length. The mouth is of the same length as the stomach, and has its margin arranged 
in a complicated series of folds. The gonads (male) extend along the whole length of 
the stomach, forming a close series of deep folds (fig. 5). This folding of the lateral 
walls of the stomach frequently gives rise to small pockets, which are probably 
receptacles for the digestion of food. 

The principal tentacles (fig. 4) are very numerous (several hundred), closely packed 
together round the margin of the umbrella. Between every two tentacles there is 
usually a very small tentacle, somewhat similar to the large tentacles in shape, but not 
so fine and slender as a typical cirrus. 


Between every large and small tentacle there is generally a long cordylus, commonly 
called a sensory club (figs. 2 and 3). On the inner side of the basal bulbs of the 
large tentacles, and very close to the velum, there is usually one black ocellus, but 
it is frequently broken up into two or three smaller pigment spots. 

Linko has found in Staurophora arctica minute ectodermal sensory vesicles, which 
are situated above the velum at its juncture with the umbrella. I have searched for 
sensory vesicles in this specimen, and have failed to find the slightest trace of one. 
Transverse sections of the umbrellar margin show a small cavity in the ectoderm in the 
same position as Linko's sensory vesicles. This cavity, however, runs through a series 
of over three hundred thick sections, so that it is not likely to be a sense-organ with 
otoliths, but it looks more like a breakage in the ectodermal layer. 

Staurophora falklandica bears a strong resemblance to S. laciniata, L. Agassiz, 
which is found in the North Atlantic, on the coasts of North America and North 
Europe. The latter species has alternating series of long and short tentacles, but the 
difference in size is very slight, and both series have ocelli. In the Falkland specimen 
there is a considerable difference in the size between the two kinds of tentacles (fig. 4). 
The very small ones are all about the same size and are without ocelli. They have the 
appearance of rudimentary tentacles. It is rather a risky point, I admit, on which to 
base the character of a new species, as there is the probability of the small tentacles 
developing into full-sized tentacles with ocelli. 

Family Eucopid^e. 
Phialidium simplex, Browne, 1902. 

Phialidium simplex, Browne, 1902, p. 282. 

Station. — Stanley Harbour, Falkland Islands, 7th January 1903. 
A single adult specimen in bad condition of this species was found in the bottle 
containing Desmonema chierchiana. 

It was one of the new species in Mr Vallentin's collection from Stanley Harbour. 


Family Halicreid^e, Fewkes, 1882. 

Vanhotfen, 1902; Maas, 1906. 

Character of Family (Maas, 1906). — Trachomedusae with numerous tentacles 
differing in size, arranged in a single row ; with eight very broad radial canals ; with a 
thick umbrella often provided with outgrowths (with a wide, thin-walled tubular 


Genus Halicreas, Fewkes, 1882. 

Generic Character (Maas, 1905). — Trachomedusse with very numerous tentacles 
(more than fifteen in an octant), which by their unequal sizes show different times of 
appearing ; with a long tubular mouth ; with eight perradial outgrowths of jelly near 
the margin of the umbrella. 


Halicreas 'papillosum, Vanhoffen, var. antarcticum, nov. 

Halicreas papillosum, Vanhoffen, 1 902, p. 68, Taf. ix. fig. 7-8, Taf. xi. fig. 30 ; 
Maas, 1905, p. 57, Taf. x. fig. 70, Taf. xi. fig. 71. 

Station 413, lat. 72° 02' S., long. 23° 40' W. Vertical net, 0-1000 fathoms. 
15th March 1904. 

Two specimens belonging to the genus Halicreas were taken at this station, and 
both are unfortunately in very bad condition. 

Specimen No. 1.— The umbrella is about 30 mm. in diameter, and is quite flattened 
out. On the ex-umbrella, at a short distance above the margin, there are eight clusters 
of papillate processes in the radii of the radial canals. In shape and size these clusters 
resemble the papillate processes of Halicreas papillosum (Vanhoffen, 1902, Taf. ix. 
fig. 7-8). There are, however, one or two isolated processes in most of the radii, just 
above the principal cluster. 

The stomach has been completely torn out, and not even a trace of it remains. The 
radial canals have nearly all shared the fate of the stomach, but two have been fairly 
well preserved from destruction, and are very broad. There are indications of six 
more canals. 

Only the lower parts of the gonads remain on four of the radial canals. The gonads 
lie in groove-like depressions of the sub- umbrella, and do not extend to the margin of 
the umbrella. The radial canals and gonads are whitish in formalin. On the surface 
of the sub-umbrella, somewhere about the periphery of the stomach, there are eight 
semi-globular projecting lumps of jelly, one between every two radial canals. 

The margin of the umbrella is in bad condition, and only the basal stumps of the 
tentacles, partly embedded in the margin of the umbrella, now remain. The stumps 
show that the tentacles are very numerous. 

Specimen No. 2. — This specimen is much smaller than the one described above, and 
looks like a contorted mass of jelly. It shows the eight clusters of papillate processes 
near the margin of the umbrella, and the eight semi-globular projections on the sub- 
umbrella. It is, however, an important specimen, because it shows that the top of the 
umbrella is capped by a small, low, somewhat spherical lump of jelly, which is studded 
with about nine small, somewhat conical processes, or tubercles. They closely resemble 
the marginal processes in shape and size, but are more isolated, being spread over a 
larger area. In the largest specimen a rather oval cap with about three or more 


processes is delineated on the top of the umbrella, which has been squeezed flat by- 
pressure in the net. 

The first specimens of this genus were described by Fewkes, 1882, under the name 
of Halicreas minimum. These were taken in the North Atlantic off the coast of the 
United States, between lat. 38°-39° N. and long. 68°-71° W., and at a consider- 
able depth. As all the specimens were in very bad condition, Fewkes's description 
and figures are consequently of little scientific value. He succeeded, however, in 
establishing the genus, but not the species. 

It was left to Vanhoffen to give the first adequate description of a Halicreas 
(H. papillosum), which he illustrated by beautiful figures. His specimens were taken 
on the voyage of the Valdivia, at many stations off the western coast of Africa 
between the Canary Islands and Cape Town, and the species was again found widely 
distributed over the warm regions of the Indian Ocean. Its chief habitat is apparently 
the tropical regions of the Atlantic and Indian Oceans, and although once taken at 
300 fathoms, it evidently prefers to live below 500 fathoms. In the Atlantic 
H. papillosum was not found north of lat. 25° N., and its southernmost range in both 
oceans was not beyond lat. 35° S. 

The geographical range of this species is important, and the southern boundary 
appears to have been fairly well fixed by the Valdivia. This ship, after leaving Cape 
Town, went to Bouvet Island, and then as far south as lat. 64°, off Enderby Land, 
returning across the Indian Ocean, via Kerguelen and New Amsterdam, to Sumatra. 
Although nets were used at many stations and at great depths, yet this species was not 
once found south of lat. 35° S. 

The Scotia specimen was taken in lat. 72°, not far off the Antarctic continent, 
which is very far outside the geographical range of H. papillosum. It is mainly for 
the sake of geographical distribution that I have decided to establish a new variety, 
which may attain the rank of a species when better specimens have been obtained. At 
present the Scotia specimens can only be distinguished from those of Vanhoffen by 
the presence of a group of tubercles on the summit of the umbrella. 

Maas has also described and figured Halicreas papillosum from specimens obtained 
by the Siboga expedition to the East Indian Archipelago. Some of his specimens have 
a cone-shaped process on the summit of the umbrella, in which they agree .with 
Vanhoffen's H. papillosum ; but some have an ordinary plain rounded umbrella, 
without any ornamentation on the summit. These latter agree with Fewkes' 
description of H. minimum, but Fewkes' specimens were in worse condition than 
those of the Scotia. 

Vanhoffen has placed in the genus Halicreas two other species (H. glabrum and 
H. rotundatum) which have not got the marginal groups of tubercles. Maas' 
definition of the genus would exclude these two species, as he restricts the genus to 
species with marginal tubercles. I certainly agree with Maas in regarding the 
marginal tubercles as a generic character. It is most probable that more than one 


species of Halicreas will be ultimately found to possess this character. I cannot call 
to mind any other kind of medusa possessing tubercles in a similar position, and they 
are certainly a most useful aid in the identification of specimens in bad condition. 

Botrynema, novum genus. 

Generic Character. — Trachomedusse with sixteen groups of tentacles (two groups 
containing many tentacles in a single row in each octant) and eight solitary perradial 
tentacles (without perradial tubercles or outgrowths of jelly near the margin of the ex- 
umbrella) . 

The characteristic feature of this new genus is the arrangement of the tentacles into 
distinct groups or clusters ; hence the generic name. 1 think less confusion will be 
produced by introducing a new genus based upon the above character, than by trying 
to emend an old genus and regarding this character as a specific one. The placing of 
the new genus Botrynema in the family Halicreidse, as emended by Maas, is 
provisional. Until better specimens showing the shape of the tentacles and the 
structure of the sense-organs can be obtained, the position of this genus among the 
Trachomedusse must remain doubtful. 

Botrynema brucei, nova species. (Plate I. figs. 8-9 ; Plate II. fig. 1.) 

Station 301, lat. 64° 48' S., long. 44° 26' W. Trawl, 0-2485 fathoms. 13th 
March 1903. 

The collection contains only one specimen which is in fairly good condition, except 
for the margin of the umbrella. 

Description. — The umbrella is very thick, about as broad as high, 25 rnm., with a 
conspicuous conical projection on its summit. The velum is very broad. The stomach 
is circular, about 9 mm. in diameter, and its centre is partly filled with a semi-globular 
projecting mass of jelly. The mouth is large and circular, owing to the thin wall of 
stomach being contracted back. There are eight broad radial canals, and a very broad 
circular canal. The gonads are shield-shaped and occupy the central part of all the 
radial canals. They are in rather a macerated and torn condition, and have the 
appearance of immature male gonads. 

The tentacles (fig. 1) have all been broken off close to the margin of the umbrella, 
but their basal ends remain partly embedded in the ex-umbrella. There are eight 
percanal (perradial) solitary tentacles and sixteen groups of tentacles. The eight 
solitary percanal tentacles are in a deep groove or furrow, which extends about 2 mm. 
over the margin of the umbrella. These tentacles are probably directed upwards, and 
those in the groups directed downwards. Halicreas papillosum has eight percanal 
tentacles which are very much larger (though similar in structure) than the tentacles 
which line the margin of the umbrella between the radial canals. 


The margin of the umbrella of Botrynema is divided into sixteen small lobes which 
are separated by furrows. There are eight percanal furrows, in which are situated the 
solitary tentacles, and eight inter-canal furrows, smaller in size and without tentacles. 
Upon each of the sixteen lobes are attached the basal ends of a group of tentacles, 
about twelve in number, arranged in a single row and in a definite order of growth. The 
basal ends of the tentacles vary in length and in size, indicating that the tentacles are 
of different lengths, which is no doubt due to differences in age. In fig. 1 the longest 
basal ends are nearest to the percanal tentacle, but in some of the other groups the 
shortest basal end occupies this position. It was in only one of the intervals between 
the groups of tentacles that very minute basal ends of tentacles were seen. They are 
indicated in the figure, and are evidently of much later growth than the tentacles on 
the lobes. Gossea corynetes (Petasidse) has eight distinct groups of tentacles, and in 
the fully grown adult a few minute tentacles make their appearance, long after the 
other tentacles, in the vacant space between the groups. 

The margin of the umbrella was carefully searched over for sense-organs, and 
none were seen ; but the margin is by no means in good condition, so that the 
absence of sense-organs is not surprising. The specimen is in formalin, and the 
colour of the principal organs is whitish. There is no trace of any red or reddish-brown 

I have much pleasure in naming this species after Dr William S. Bruce, the leader 
of the Scottish National Antarctic Expedition. 



Family AtolliDjE. 

Atolla chuni, Vanhoffen, 1902. 
Atolla chuni, Vanhoffen, 1902, p. 12, Taf. i. fig. 1-2, Taf. v. fig. 26. 

Station 450, lat. 48° S., long. 9° 50' W. Trawl, 1332 fathoms. 12th April 1904. 

The collection contains one specimen, which evidently got damaged in the trawl. 
The mouth, stomach, and gonads are partly torn away, but the margin of the umbrella 
is in fairly good condition. 

This species was first found by the Valdivia expedition, and has been beautifully 
figured by Vanhoffen. Two specimens were taken by the Valdivia between the 
Cape of Good Hope and Bouvet Island (lat. 42° S., long. 14° E.) in a vertical 
net at 1500 metres (821 fathoms), in the Antarctic current with a temperature 
of0°-8 0. 

Atolla chuni is distinguished from the other species of the genus by the presence 


of warts (Gallertperlen) on the marginal lobes. In the Scotia specimen the warts are 
quite conspicuous upon the lobes. Some of the lobes have nine warts (four on each 
side and one in the middle), agreeing in number, position, and size with Vanhoffen's 
description. Many lobes, however, have only seven warts (three on each side and one 
in the middle), and as the two smaller uppermost warts are the missing ones, these 
lobes are probably at an earlier stage in development. The central wart was missing 
in two lobes — perhaps an instance of variation. 

There are twenty-four tentacles and twenty-four sense-organs. The gonads are 
completely torn away in places, and only two perfect and two imperfect sacs remain. 
There is sufficient evidence to show that the genital sacs form very nearly a complete 
ring. Two of the sacs are close enough to touch each other, and others show only a 
very slight separation. 

Measurements. — (The specimen is rather flattened out.) Diameter of the central 
disc, 23 mm. Diameter of the muscle-band, 35 mm. Diameter of the stomach, 
16 mm. Length and width of pedalia, 3*5 mm. 

Atolla ivyvillei, Haeckel, 1880. 

Atolla wyvillei, Haeckel, 1880, p. 488; Haeckel, 1881, p. 113, pi. xxix. ; Vanhoffen, 1902, p. 13, 
Taf. v. fig. 22. 

Station 413, lat. 72° 02' S., long. 23° 40' W. Vertical net, 0-1000 fathoms. 
15th March 1904. 

At this station was taken an Atolla which is in very bad condition. Its stomach 
and gonads are completely gone, and the umbrella is flattened out. 

The periphery of the central disc is distinctly divided into large lobes, and the 
radial furrows separating the lobes are rather broad. The furrows resemble those in 
Vanhoffen's figure of Atolla ivyvillei much more than those in the one he gives of 
A. ver villi. It is upon the shape of the furrows that I have decided to call the 
specimen A. wyvillei rather than A. verrilli. There are twenty-two tentacles, and 
the marginal lobes are without warts. The pedalia have a slight longitudinal 
furrow. As the diameter of the muscle-band is only 25 mm., the specimen is about 

Atolla ivyvillei was first found by the Challenger expedition at the depth of about 
2000 fathoms. Three specimens were taken about midway between the Kerguelen 
Islands and Melbourne (lat. 53° S., long. 108° E.), and two specimens in the South 
Atlantic Ocean off the coast of Patagonia (lat. 42° S., long. 56° W.). The Valdivia 
expedition secured a specimen south-east of Bouvet Island (lat. 56° S., long. 14° E.) 
in a vertical net down to 1500 metres (821 fathoms), and another specimen 
off Kemp Land (lat. 63° S., long. 57° E.) in a dredge down to 4600 metres 
(2517 fathoms). 




Family Pelagiid^e. 

Pelagia perla (Slabber), 1775. 

Station 542, lat, 37° 56' N., long. 29° 11' W. 4th July 1904. 

Station 543, Fayal Harbour, Azores. Surface. 5th July 1904. 

Station 544, lat. 39° 15' N., long. 26° 55' W. (north of the Azores). Surface. 
7th July 1904. 

Fourteen specimens were taken in Fayal Harbour. The smallest is about 50 mm. 
in diameter, and the largest about 75 mm. The umbrella is about twice as broad as 
high, and externally covered with large elliptical warts. The marginal lobes are 
quadrilateral, with rounded corners, a little broader than long. The oral arms are about 
twice the length of the radius of the umbrella, and the oesophagus about two-thirds the 
length of the radius. 

Eighteen specimens were collected at the station north of the Azores. They are all 
young stages, about 20-30 mm. in diameter, and have roundish warts on the ex- 
umbrella. At Station 542 a single specimen in very bad condition was taken. 

There is probably only one species of Pelagia in the North Atlantic, though six 
have been described. 

Family Cyaneid^e. 
Desmonema, L. Agassiz, 1862. 

This genus was established by L. Agassiz for the reception of Chrysaora gaudi- 
chaudi, Lesson, which was first discovered during the voyage of the Coquille in 
Soledad Bay, Falkland Islands, and also near Cape Horn. At the same time Agassiz 
made another new genus called Couthouyia, of which only a very brief account was 
given. Haeckel emended the definition of the genus Desmonema so as to include 
Couthouyia, and added three species, namely, Couthouyia pendula, L. Agassiz ; 
Cyanea imporcata, Norman ; and a new species called Desmonema annasethe. 
According to Haeckel's definition, Desmonema differs from Cyanea in having the 
tentacles arranged in a single row instead of in several rows. 

Vanhoffen a little later also emended the genus, and rejected Haeckel's distinction 
as to the arrangement of the tentacles, on the ground that Cyanea passes through a 
stage with the tentacles arranged in a single row. Vanhoffen's definition of the genus 
Desmonema, is based upon Agassiz's brief description of Couthouyia pendula, and the 
genus therefore should have been called Couthouyia. According to Vanhoffen, 
Desmonema differs from Cyanea in possessing twenty-four marginal lobes. In Desmo- 
nema the tentacular lobes are not divided as in Cyanea. The gonads are shorter 
and do not hang down so far as those of Cyanea, and the oral arms taper from a broad 


base. Vanhoffen places three species in the genus, the two old ones, D. pendula 
from Tierra del Fuego, D. gaudichaudi from the Falklands and from near Cape Horn, 
and a new species called D. chierchiana found at Punta Arenas in the Straits of 
Magellan. This last species has been fully described and well figured, but I am able 
to add a further account after the examination of five specimens in different stages of 
development found in Stanley Harbour, Falkland Islands. It will be noticed that the 
above three species are inhabiting the same region, and I have but little doubt that 
they all belong to one species, which, in accordance with the law of priority, should be 
called Desmonema gaudichaudi (Lesson). 

The description of Chrysaora gaudichaudi by Lesson is incomplete, and is based 
upon an abnormal specimen with twelve groups of tentacles instead of the normal 
eight. Consequently the shape and size of the marginal lobes would differ from those 
of the normal type. Agassiz's description of Couthouyia pendula is practically a 
generic one, but he does state that the oral arms are of an extraordinary length. 

Vanhoffen points out that Desmonema, chierchiana is distinguished from 
D. gaudichaudi by shorter oral arms. The largest specimen in the Scotia collection 
shows that D. chierchiana has much longer arms than those mentioned by Vanhoffen. 
The length and size of the oral arms are of little use for a specific character, on account 
of their great power of expansion and contraction. Vanhoffen also points out that 
the shape and size of the marginal lobes of D. chierchiana differ from those of 
D. gaudichaudi. It must be remembered that Lesson's specimen was an abnormal 
one. The marginal lobes of the Scotia specimens are not quite like those of 
Vanhoffen's Punta Arenas specimens, the difference being probably due to variation 
and not to a different geographical race. 

Vanhoffen lays a certain amount of stress on the gonads of Desmonema being 
smaller than those of Cyanea, and not hanging down so far. The largest specimen in 
the Scotia collection has much larger gonads than Vanhoffen's specimen, and they are 
just like the gonads of Cyanea capillata. 

I rather think that the specific character of D. chierchiana lies in the arrangement 
of the tentacles in practically a straight row, adjacent to the distal edge of the 
circular muscles. Concomitant with the straight row of tentacles are the undivided 
tentacular lobes, and the absence of radial muscles in the marginal lobes. The 
tapering of the oral arms may also come in as a useful aid for the determination of 
the species. The ultimate fate of Desmonema as an independent genus depends upon 
the revision of the genera and species of the Cyaneidse. Maas (1904) has also com- 
mented upon the relationship of Desmonema and Cyanea, and the necessity for a 
revision of the species. 


Desmonema chierchiana, Vanhoffen, 1888. (Plate II. fig. 2.) 

Desmonema chierchiana, Vanhoffen, 1888, p. 17, Taf. i. fig. 4. 
? Chrysaora gaudichaudi, Lesson, 1830, p. 114, pi. xiii. fig. 1. 
? Desmonema gaudichaudi, L. Agassiz, 1862, vol. iv. p. 166. 
1 Desmonema gaudichaudi, Haeckel, 1880, p. 527. 
1 Coutliomjia pendula, L. Agassiz, 1862, vol. iv. p. 118, p. 163. 
% Desmonema pendula, Haeckel, 1880, p. 528. 

Station. — Stanley Harbour, Falkland Islands, 7th January 1903. 

The Scotia brought home three specimens, the largest of which is in splendid 
condition, but the other two are not quite perfect. The large collection of medusas 
made by Mr Rupert Vallentin in Stanley Harbour during the summer of 1898-1899 
is still in my possession, and it contains two early stages of this species in excellent 
condition. At the time when I examined Mr Vallentin's collection I was unable to 
identify these young stages, as they were just too young to show clearly the generic 
characters. By the aid of the Scotia specimens I have now been able to identify them, 
and the five specimens form a nice series showing different stages in development. 

Specimen A (Vallentin Coll.). — This specimen is the youngest of the series, and 
its development has not proceeded very far beyond the Ephyra stage. 

The umbrella is very thin and flat, measuring about 14 mm. in diameter. The ex- 
umbrella has scattered over its surface numerous small clusters of nematocysts, which 
are most prominent and conspicuous on the marginal lobes. The stomach is circular, 
about 9 mm. in diameter, and has sixteen radiating pouches which are separated from 
each other by radial septa. Eight of the pouches are opposite sense-organs, and eight 
opposite the groups of tentacles. Within the stomach are four interradial bundles of 
gastric filaments. From the mouth hang down four oral arms, which are a little 
longer than the radius of the umbrella. They have the appearance of four large lips 
folded along the centre. Their external surface is closely covered with small warts 
containing nematocysts. 

There are eight adradial groups of tentacles, each group containing one long 
tentacle and four to six minute tentacles or tentacular buds. The presence of one long 
tentacle in each group indicates that in an earlier stage only eight tentacles are present. 
The tentacles are within the margin of the umbrella, adjacent to the distal edge of the 
gastral pouches. There are eight tentaculocysts (rhopalia), four perradial and four 

The margin of the umbrella is divided into sixteen lobes or lappets by eight deep 
clefts leading to the sense-organs, and by eight clefts, not quite so deep, which occur in 
the middle of the tentacular lobes (the lobes opposite the groups of tentacles). At this 
stage, however, each marginal lobe is composed of one ocular and half a tentacular lobe. 
Later on this cleft in the middle of the tentacular lobes becomes obliterated. The 
marginal lobes have a definite series of canals connected with the pouches of the 
stomach. The canal system will be described in the next stage. 


Specimen B (Vallentin Coll.). — The umbrella is fairly thin and flat, measuring 
about 25 mm. in diameter, and is very pellucid in formalin. 

The stomach is about 15 mm. in diameter, with sixteen radial pouches. The 
tentacular pouches are a little broader than the ocular ones. The oral arms are about 
twice as long as the radius of the umbrella. 

The tentacles show an increase in number and in size. In each group the central 
tentacle is very conspicuous by its great length (fig. 2). It is at least six times as 
long as the other tentacles. Each group contains about three to six tentacles and 
about six tentacular buds. 

The cleft in the middle of the tentacular lobe has nearly disappeared (fig. 2), and 
the ocular lobes are just marked off by a slight incurving of the margin. Each ocular 
lobe has a broad canal which gives off a number of short lateral branches, which in this 
stage are just beginning to develop. The tentacular lobes have a number of canals, 
one between every two tentacles. These canals in the later stages have each a series 
of bifurcations. 

Specimen C [Scotia Coll.). — The umbrella is moderately curved and is fairly thick, 
measuring 50 mm. in width and 20 mm. in height. 

The stomach is circular, about 25 mm. in diameter, with sixteen radiating pouches. 
The width of the tentacular pouches at their distal end is about 10 mm., and the ocular 
pouches about 7 mm. The oral arms are in a damaged condition. The width between 
the pillars of the arms, measured across the oval where the gonads are just beginning 
to show, is 11 mm., and the width of the pillar is 3 mm. at its narrowest point. 

There are eight groups of tentacles arranged in a straight single row. Each group 
consists of twelve to fifteen tentacles, with the largest in the middle. The large 
tentacles are all in a broken condition. 

The cleft in the middle of the tentacular lobes has practically disappeared. The 
ocular lobes are slightly more prominent, projecting further out than the tentacular 
lobes. The canal system is of the same pattern as in specimen B, but the branches 
have more bifurcations. 

Specimen D. — The umbrella is fairly thick, about 80 mm. in width and 40 mm. in 
height. Clusters of nematocysts are closely scattered over the ex-umbrella and the 
marginal lobes, but the warts are absent, so that the surface looks quite smooth. 

The stomach measures about 35 mm. in diameter. The width of the tentacular 
pouches at the distal end is about 18 mm., and the ocular about 12 mm. This speci- 
men is in a damaged condition, having lost its oral arms, gonads, and tentacles. The 
distance between the pillars of the oral arms is about 15 mm. There are indications of 
twelve to fourteen tentacles in each of the eight groups. 

The ocular and tentacular lobes are united into one large lobe with an even margin. 
The marginal distinction between the lobes has become obliterated, and the medusa has 
the appearance of possessing eight broad marginal lobes which are separated from each 
other by the deep ocular clefts. But owing to the incurving of the margin of the 


umbrella its margin appears to be deeply lobated. Upon the periphery of the ex- 
umbrella there are sixteen deep radial furrows, in line with the sixteen radial septa 
which divide the pouches of the stomach. It is along these furrows that the folding 
takes place, dividing the margin into sixteen broad folds, each of which is again sub- 
divided by the eight ocular clefts, and by an indentation in the middle of the tentacular 
lobes. On straightening out the margin of the umbrella all these folds disappear, and 
one then sees the true lobation of the margin. 

Specimen E. — This specimen is the largest in the Scotia collection, and as it is in 
excellent condition full details are given. 

The umbrella is rather flat, and has probably flattened out in the jar owing to the 
pliability of the jelly. It measures about 160 mm. in diameter with the marginal 
lobes folded inwards. For the size of the umbrella the jelly is not very thick. 

The stomach is circular in outline, about 75 mm. in diameter, with sixteen radial 
pouches. The tentacular pouches are 35-40 mm. in width at their distal margin, and 
the ocular about 25-30 mm. The circular muscle-bands of the sub-umbrella are upon 
the lower wall of the pouches, and are separated radially by sixteen furrows which 
correspond in position to the septa of the gastral pouches. I have not been able to 
find any radial muscle-bands on the sub-umbrella or the marginal lobes. 

The four oral arms are about 350 mm. in length, and are probably capable of 
extending to a much greater length. Along the whole length of each arm runs a 
double frill. These frills are very broad near the mouth and gradually decrease in 
width towards the distal end of the arms, and give to the arms a long, slender appear- 
ance. The width of the pillars of the arms is 7-8 mm. across the narrowest part. 

The four gonads are similar in position and shape to those of Cyanea capillata. 
They hang down in clusters from between the bases of the oral arms, and are much 
larger than the gonads of the specimen figured by Vanhoffen. As the medusa lies 
flat in the dish the gonads extend about half way across the circular muscles. The ova 
are quite small and immature, so that in a fully ripe condition the gonads should be 
much larger in size. The genital openings between the oral arms are oval in shape and 
measure nearly 40 mm. in length. 

There are eight groups of tentacles arranged in a straight single row, adjacent to 
the outer edge of the circular muscles. The largest tentacles are in the middle of the 
group and the smallest on either side. They are like long threads with a dark reddish 
endoderm, and externally covered closely with roundish warts containing nematocysts. 
The number of tentacles in each group is as follows: 20, 21, 19, 17, 18, 20, 21, 19. 

The eight sense-organs are quite simple in their construction. The tentaculocyst 
in the early stages lies in a rather broad U-shaped groove, which is partly covered on 
the sub-umbrella side by the margins of the two ocular lobes. In the adult stage, 
owing to the increase in the size of the ocular lobes, the groove is longer and is 
completely covered by the ocular lobes. A canal from the pouch of the stomach runs in 
the roof of the groove, and at its distal end is situated a small tentaculocyst, about twice 


as long as broad, containing concretions. On the wall of this canal near the tentaculo- 
cyst a small roundish swelling is present in the earlier stages, and a long ridge-like 
swelling in the adult. It is probably a sensory pad. 

The medusa as it lies on its back in a basin, with the margin of the umbrella folded 
inwards, shows very clearly the sixteen radial furrows on the periphery of the ex- 
umbrella. These furrows are lines of weakness, for they reduce the thickness of the 
marginal jelly. In two places the ocular lobe is completely isolated from the tentacular 
lobe, the tearing asunder taking place along the furrow. The folding of the incurved 
margin is similar to that described for specimen D. 

The tentacular lobes measure 55 mm. in breadth, and are about two and a half times 
as broad as the ocular lobes. Their length is about 30 mm. In this specimen the 
length of the tentacular and ocular lobes is about the same. In Vanhoffen's specimen 
the ocular lobes are a little shorter than the tentacular. In most of the octants there 
is no indentation between the tentacular and ocular lobes ; an even margin extends 
from one ocular cleft to another. But in two places there are natural indentations 
between the ocular and tentacular lobes, very much like the indentations drawn by 
Vanhoffen. Evidently there is a fair range of variation in the shape and size of the 
marginal lobes, and taken by themselves they are not a specific character. The 
marginal lobes and the surface of the ex- umbrella are quite smooth. Apparently the 
warts conspicuous on the lobes in the early stages disappear before the medusa reaches 
the adult state. 

The arrangement of the canals in the marginal lobes is similar to that in the earlier 
stages, but there are more bifurcations. The canal system has been well illustrated 
in Vanhoffen's figure. 

In Mr Vallentin's manuscript notes I have found several references to Desmonema 
in Stanley Harbour, from November 1898 to February 1899. During November large 
and small specimens were very abundant, the largest being about 1 foot (300 mm.) 
in diameter. At the end of December specimens up to 2 feet (600 mm.) in diameter 
were abundant. Records show that large specimens were present in considerable 
numbers during January and February. On 6th February thousands were stranded 
on the shore after a strong wind. 

Family Ulmarid^e. 

Phacellophora ornata (Verrill, 1869). (Plate II. figs. 3-4). 

Callinema ornata, Verrill, 1869, p. 116. 
Callinema ornata, Fewkes, 1888, p. 234, pi. vi. 
Phacellophora ornata, Haeckel, 1880, p. 643. 
Phacellophora ornata, Vanhoffen, 1906, p. 59, figs. 25-26. 

Station 98, lat. 34° 2' S., long. 49° 7' W. Surface. 28th December 1902. 
A single specimen was taken at this station, which is about five degrees east of 
Montevideo. It is unfortunately damaged, having completely lost the oral arms. 


Iii the Voyage of the Scotia, a book which gives an interesting and popular account 
of the work done by the Scottish National Antarctic Expedition, there is an allusion 
to this medusa : "On the 28th, in 33° 51/ S., 48° 48' W., we crossed a great yellow 
band of gelatinous scum stretching from horizon to horizon. It proved to consist of 
microscopic algse (Desmidse) closely allied to the diatoms, and in the scum were numbers 
of Portuguese men-of-war (Physalia), Jelly-fish (Aurelia), Ctenophores, and other 
organisms." The specimen was labelled Aurelia sp. 

Description. — The umbrella has flattened out, and as it lies in the dish it is about 
five times as broad as high. The diameter, measured across to the circular canal, is 
about 1 20 mm. The central portion of the ex-umbrella is slightly raised and is covered 
with small warts, which are quite visible to the naked eye. The periphery of the 
umbrella is quite smooth. The base of the stomach is small for the size of the 
umbrella, measuring only 30 mm. in diameter. The oral arms have disappeared, and 
only broken stumps remain. 

From the stomach to the circular canal run forty-two broad radial canals. The 
canals in the radii of the sense-organs are branched ; the branching is rather irregular, 
and there is a tendency to anastomosis (fig. 3). The canals in the radii of the tentacles 
are unbranched, with somewhat irregular outlines. 

The margin of the umbrella is not quite perfect, but nine sense-organs are present, 
and the number of branched radial canals indicates that two sense-organs are missing. 
The tentacles are arranged in a single row along the outer edge of the circular canal. 
The row is broken by the sense-organs so that groups are formed. 

The character of the genus Phacellophora is the presence of sixteen sense-organs, 
by which it is distinguished from Sthenonia with eight sense-organs. The Scotia 
specimen has eleven sense-organs and eleven groups of tentacles, and it is evident from 
the irregular distribution of the sense-organs that the specimen is an abnormal one. 
In two places on the margin the sense-organs are very close together, with just a few 
tentacles between them. 

There are twenty tentacles in the largest group, and a very small tentacular bud 
lies between every two tentacles. The tentacles are all about the same size and have 
a well-marked character. They have the appearance of long, tapering threads of clear 
jelly, with a thin whitish line along the outer side. On applying higher magnification 
one sees that the tentacle has a canal running along its whole length. This canal in 
the basal portion of the tentacle is very large, forming a large hollow cavity ; but it 
quickly decreases in size until it becomes a very tiny tube, which is perhaps solid 
towards the distal end, running not in the centre of the tentacle, but close against the 
outer side (fig. 4).. It is only the basal portion of the tentacle that is really hollow, 
and the canal is a rudiment of an axial endodermal cavity. Along the outer side of 
the tentacle runs a double row of warts, which are loaded with nematocysts. There 
are no muscle-fibres visible in the tentacles. 

The ocular lobes (fig. 3) on the margin of the umbrella are thick, being flat on the 


inner side and rounded on the ex-umhrellar side. They are about 12 mm. in length 
and 6 mm. in width, having the outer edge with rounded corners, and are separated 
from the tentacular lobe by a slight marginal cleft. Along the centre of each ocular 
lobe runs an unbranched canal. 

The tentacular lobes are about as long as the ocular lobes, and have an undulating 
margin. These lobes correspond in number and width with the groups of tentacles. 
Each lobe has a few isolated canals which are in connection with the circular canal. 
Some of these canals are slightly bifurcated at the distal end. 

This specimen agrees very well with Fewkes' description and figures of Callinema 
ornata so far as the canal system, sense-organs, tentacles, and marginal lobes are 
concerned. Phacellophora ornata up till now has only been taken in one locality, 
namely, at Eastport, on the coast of Maine, U.S., in the North Atlantic. Verrill 
obtained three specimens, and many years later Fewkes secured another. 

Aurelia solida, Browne. 
Aurelia solida, Browne, 1905, p. 960, pi. xciv. 

Station 539, lat. 33° 53' N., long. 32° 27' W. Surface. 1st July 1904. 

Four specimens were taken at this station, which is about 15 degrees west of 
Madeira. Two are in very good condition, and two have a dilapidated umbrellar 
margin. My original description of Aurelia solida was based upon a single specimen 
obtained by Mr Stanley Gardiner during his expedition to the Maldive and Laccadive 
Archipelagoes. I was certainly surprised to find this species from a mid-North Atlantic 
station in the Scotia collection. The occurrence of this Aurelia in the Indian Ocean 
and the North Atlantic shows a very wide geographical range, and it is most probable 
that the species has been taken long ago and recorded under another name. But none 
of the descriptions and figures of the numerous species of Aurelia show the characters 
of this medusa. It is certainly not Aurelia aurita, neither can it be regarded as a 
variety of that species. 

Description. — The umbrella is hemispherical, about twice as broad as high, and 
thick. The four specimens are about the same size, 55-60 mm. in width, and smaller 
than the Maldive specimen, which measured 80 mm. in width. The ex-umbrella is 
covered with small circular clusters of nematocysts. 

The mouth is cross-shaped and formed by four thick, somewhat triangular 
segments, which are about 20 mm. in length. The four segments are interradial and 
separated from one another. The mouth has the appearance of a cone-shaped lump of 
jelly cut longitudinally into four equal segments, and the four pieces placed together 
again. Along the outer edge of each segment runs a short frill, which at j the base 
grows out to form the oral arm, which extends a little way beyond the margin of the 
umbrella. ,., •.- •.- .,.-..-, ...,.,. 



The radial canals show a certain amount of variation, especially in the interradial 
series, and a slight amount of anastomosis takes place near the margin of the umbrella. 

The sense-organs are exactly like those described in the Maldive specimen. The 
tentaculocysts are at right angles to the margin of the umbrella, pointing towards the 
ex-umbrella, and stand in a kind of alcove which is at the end of a deep groove open on 
the dorsal or ex-umbrellar side. It is by the position and structure of the sense-organ 
that this species can at once be distinguished from Aurelia aurita. 

The subgenital cavities are nearly circular, and have a fairly thick covering, with a 
small circular aperture about 2 mm. in diameter. They are similar in shape to those 
of the Maldive specimen, but are not so prominent and have not such a thick external 
wall. The gonads form oval or horse-shoe-shaped bands. They have not reached 

Jelly-fishes in an unrecognisable condition were taken at the following stations : — 

Station 325, Scotia Bay, South Orkneys, 25th March 1903. "Fragments of a 

Station 414, lat. 71° 50' S., long. 23° 30' W. ; 0-1000 fathoms. Vertical net. 
One Trachomedusa. Fragments of the umbrella of a Scyphomedusa. One specimen of 
a large Pleurobrachia ? 

Station 415, lat. 71° 28' S., long. 22° 32' W. ; 0-2338 fathoms. A lump of 
jelly found in a sounding-tube. 


Agassiz, L., 1862, Contributions to the Natural History of the United States of America, vol. iv. Boston. 
Browne, E. T., 1902, "A Preliminary Report on Hydroniedusae from the Falkland Islands," Ann. Mag. 

Nat. Hist., ser. 7, vol. ix., pp. 272-284. 
,, 1905, " Scyphomedusae " in Gardiner's Fauna and Geography of the Maldive and Laccadive 

Archipelagoes, vol. ii., pp. 958-971, pi. xciv. Cambridge. 
,, 1907, "A Revision of the Medusae belonging to the Family Laodiceidae," Ann. Mag. Nat. 

Hist., ser. 7, vol. xx , pp. 457-480., J. W., 1888, "On certain Medusae from New England," Bull. Mus. Cornp. Zovl., vol. xiii., 

pp. 209-240, pis. i.-vi. 
Haeckel, E., 1879-80, Das System der Medusen. Jena. 

„ 1881, "Report on the Deep-sea Medusae dredged by H.M.S. Challenger during the years 

1872-1876," Challenger Reports, vol. iv. 
Lesson, R. R, 1830, Voyage autour du Monde . . . sur la corvette "La Coguille" pendant les annees 

1822-1825: Zoologie. Paris. 
Maas, O., 1904, " M^duses provenant des campagnes des yachts Hirondelle et Princesse Alice " (1886-1903), 

Result. Camp, scient. Monaco, fasc. xxviii., 71 pp., 6 pis. 
„ 1905, "Die craspedoten Medusen der Siboga-Expedition," Monograph X., 84 pp., 14 Tafn. 

„ 1906, "Die arktischen Medusen (ausschliesslich der Polypomedusen)," Fauna Arctica, Bd. iv., 

pp. 481-526. 
„ 1906, Expedition antarctique Beige " Belgica" : Zoology, "Medusen," 32 pp., 3 pis. Anvers. 


Vanhopfkn, E., 1888, " Untersuchungen uber semaeostome und rhizostome Medusen," Bibliotheca Zoologica, 

Heft iii., 54 pp., 6 Tafn. 
„ 1902, "Die acraspeden Medusen der deutschen TiefseeExpedition 1898-1899," Wissensch. 

Ergebnisse ..." Valdivia," Bd. iii., pp. 1-52, Taf. i.-viii. 
,, 1906, "Acraspede Medusen," Nordischen Plankton, part xi., pp. 40-64, 37 text figures. 

Verrill, A. E., 1869, "Description of a Remarkable new Jelly-fish and two Actinians from the Coast of 

Maine," Amer. Journ. Set., 1869, pp. 116-118. Reprinted in Ann. Mag. Nat. 
Hist., ser. 4, vol. iv., p. 161. 


Plate I. 

Staurophora falklandica, n. sp. 
Fig. 1. Oral view. Natural size. 
Fig. 2. Cordylus partly contracted. x 140. 
Fig. 3. Cordylus expanded. x 140. 

Fig. 4. Margin of the umbrella, showing the tentacles and cordyli. Inner side, x 25. cc, circular 
canal ; oc, ocellus ; sc, cordylus ; T 2 ., rudimentary tentacle ; v., velum. 

Fig. 5. Inner side of the wall of the stomach with gonads. x 10. g., gonads ; m., margin of the mouth. 

Fig. 6. Lateral view of the base of a tentacle. x 25. 

Fig. 7. Basal bulb of a tentacle. Outer side. x 25. t.r., the spur or root of the tentacle. 

Fig. 8. Botrynema brucei, n. g., n. sp. Lateral view. The margin of the umbrella is curled inwards, x 2. 
Fig. 9. Botrynema brucei. Diagram showing the position of the gonads. 

cc, circular canal; g, gonads; m., margin of mouth ; r., radial canal; st., stomach. 

Plate II. 

Fig. 1. Botrynema brucei. A sketch showing the arrangement of the tentacles on the margin of the 
umbrella. The velum is omitted ; its position is below the tentacles. Aboral view. x 9. per., percanal 
(perradial) radius; inter., intercanal radius; p.t., percanal tentacle. 

Fig. 2. Desmonema chierchiana, Vanhoffen. A portion of the sub-umbrella showing the arrangement 
of the circular muscles, tentacles, sense-organs, and marginal lobes in a young stage, x 3. cm., circular muscles ; 
occ, canal of the ocular lobe; ocl., ocular lobe; v., tentaculocyst (rhopalium) ; spt., septum dividing the 
pouches of the stomach ; t.c, canal of the tentacular lobe ; t.l., tentacular lobe. 

Fig. 3. Phacellophora ornata (Verrill). A portion of the sub-umbrella showing the arrangement of 
the radial canals, sense-organs, and tentacles. Natural size, cc, circular canal ; occ, canal of the ocular 
lobe; ocl., ocular lobe; ?*., tentaculocyst; r.c, radial canals; s., stomach; t., tentacles (cut off at base); 
t.c, canal of the tentacular lobe ; t.l., tentacular lobe. 

Fig. 4. Phacellophora ornata. Portion of a tentacle, showing the arrangement of the nematocysts. 
e., endodermal canal; n., nematocysts. 


On the sub-umbrella and on the gonads of the large Desmonema chierchiana many 
amphipods were found. Mrs E. W. Sexton of Plymouth kindly undertook their 
identification, and informs me that they are Hyperia gaudichaudi (H. Milne- 
Edwards). Most of the specimens were immature, but adult males and females were 
present. Hyperia gaudichaudi has been taken off the coast of Chile, and was present 
in the Antarctic collections made by the Southern Cross and by the Discovery. 

Trans. Roy. Soc.Edm r Vol. XLVI. 

E.T.Browne: Medus/e of the Scottish National Antarctic Expedition-Plate I. 


Fig. 2. 


.<■ '.-.■ 


-■■ ■'•". 


■' ■ sT ■ ■ ■ 




Fig. 1. 

HV ?# 




■ ■ 

f ! 



--T 2 




Fig. 3. 


Fig. 4. 

Fig- 5. 

^■r' \ 

Fig. 7. 


Fig 9- , 




MrFarUne iErskine Ltth Erjn 

Trans. Roy. Soc. EdirrT Vol. XLVI. 

K.T.Browne: Medus/e of the Scottish National Antarctic Expedition — Plate II. 




£ g§* -223 | 

Fig. 4 


tL<sp> 31 &]) 



M'f'jrlane kErslunc LieVEdm'' 

>f* H "*£&> 

( 253 ) 

XL — On g-Functions and a certain Difference Operator. By the Rev. F. H. 
Jackson, M.A. Communicated by Professor Chrystal. 

(MS. received January 16, 1908. Read February 17, 1908. Issued separately April 21, 1908.) 


Part I. 

Introduction .......... 

§ 1. Definitions of A*> A*- 

§ 2. Equation A x y = y- Transformation of the Power Series 2«nX" 
§ 3. Elementary properties of the operative symbol A. Comparison with the differential 
Analogue of ~Duv = vDu + uDv, etc. ...... 

§ 4. Interpretation of A~", q- Finite Integration ..... 

§ 5. Examples of Inverse operations ...... 

§ 6. Solutions of A Equations. Function of multiplicative period q, F(qx) = F(x) 

§ 7. Forms of k n u x , & n u x v z ' . 

§ 8. Maclaurin's Series and Conjugate Form ..... 





il operator 








Part II. 

§ 1. Infinite product of which the general factor is (l+a 1 xq r + a 2 x 2 q 2r + . . . a n x"q rn ) 

§ 2. Expression of this product as a power series ........ 

§ 3. Expression of the product of n functions Q(a 1 x) Q(a 2 x) .... 0(«„a;) as Aj + A 2 cos 2x + A 3 cos 4x +• 

§4. Special Series for J acobi's function ©"(a;). ........ 

, (l+ax)(l + aqx) 2 (l+aq 2 xf .... „ 

§ 5. Expression of {l+bx){l+biX y (1 + bq 2 zf _ . .. as 2 V • 

. (1 + 2 ^ cos e + a 2 a- 2 g 2 )(l + 2(ta;g 2 cos 6 +a 2 x 2 q i )\l + '2,axq 3 cos e + a 2 x 2 q % f . , 

3 6. Expression of (Y^ bxq cos g + tftf^yi + 2bxq 2 cos e + ^^ + 2bx f cos B + b 2 xhff . 

§ 7. Coefficients /j, n , v n allied to X n . 

§ 8. Relations of these with q generalisations of Bessel's coefficients 



cos ne 






Part III. 

§ 1 . The general hypergeometric series 5-form .... 

§ 2. Special g-difference equations allied to Bessel's differential equation 

§ 3. Solutions corresponding to Hankel's Y„ 

§ 4. Solutions of the equation satisfied by J,,; . J„ . 

§ 5. Relations between solutions ...... 

§ 6. Recurrence and special A relations satisfied by the solutions . 






Part I. 


In this paper my object is, primarily, to investigate the properties of a certain 
operative symbol A", which appears to be of great utility in discussing (/-functions. 
The first part of the paper will consist of an investigation into the various forms of 

and the nature of the inverse operations symbolised by A "". With certain restrictions 
as to continuity, etc., <p{x) will denote an arbitrary function of x. In the second part 
of the paper such products as 

(1 + a 1 x + a 2 x 2 + . . . + a li x ll )(l + a 1 qx + a 2 q 2 x 2 + . . . + a n q n x n )(l + a 1 q 2 x + . . . + a n q"~"x n ) . . . . , 

< '^)( 1+ f)'( 1+ ?T 

<^>K*) 2 K?)° ' 

( 1 + 2(j xcoa6 + q 2 x 2 )( 1 + 2yV cos 9 + (/V) 2 (l + 2q*x cos 9 + q^x 2 f 

(1 +■ 2qy cos 6 + q 2 y 2 )(l + 2q 2 y cos 9 + q i y 2 ) 2 (l + 2q*y cos 9 + q 6 y 2 f ' 

= \ (x , y) + X^x , y) cos 9 + \ 2 (x ,y)cos29+ , 

will be discussed, and relations found between the coefficients X(x, y), and the q 
generalisations of Bessel's Functions denoted by the symbols J [n] (x), *$ {n] {x), in previous 

In the third section of the paper certain special A -equations will be discussed in 
connection with their limiting forms as linear differential equations. 

In the appendix to Chapter ii. of Heine's Kugelfunctionen\ there is given a 
discussion of the special (/-function <p(ci, b, c, q, x) analogous to the Hypergeometric 
series F(a, b, c, x). Heine makes use of a difference symbol A, which he defines by 

the equation 

A</>(a ,b ,c,q,x) = <j>(a ,b,c,q, qx) - <£(a ,b,c,q,x). 

It is obvious that the repetition of the operation symbolised by A gives us 

A n &(x) = ^q n x)-n^q n - 1 x) + ?^^<f>(q n -' i x)+ ( - 1)"<%) . 

Now, in q series, such coefficients as the Binomial coefficients do not usually appear 
The (/-equivalent of H is [1][2] . . . [r], in which [r] = (q r - 1 )/(</- l) ; therefore, in 
discussing (/-functions, 1 think it desirable to use an operative symbol which gives rise 
to a series in which the coefficients follow the (/-Binomial form, i.e. 

^ , ^x) = ^q n x)-[7i]^q n - 1 x) + q^[ Ti ^^(q n - 2 x)- ( - l)^'""" 1 '^) 


* Trans. Roy. Soc. Edin., vol. xli. pp. 105-118, 399-407, etc. Proc. L.M.S., series 2, vol. ii. pp. 192-221 ; vol. iii. 
pp. 1-23. 

t Heine, Kuyelfunctionen, ed. 1878, pp. 99 et seq ; also Thomje, Crelle's Journal, vol. lxx. 


We proceed as follows by defining an operation A<fi(x) as the quotient of two 

§ 1. Forms of A n . 
AWs)}-*^) -*<*>, (1) 

We see that in the limiting case, q= 1, this equation is D {$(«)} = &'(x), provided that 
<£(;») has a first derivative &(x). Repeating the operation symbolised by A, we obtain 

as{<*>(*)} = ^ x) ~ frjfo^ + <**<*> .... (2) 

(q 2 x - qx)(qx - x) 
and, by induction it is easy to establish 

A«{$( a -)}= 4>( ' /a;) ~ w * ( ' / "" lx)+ • • • • ( - 1 ) r g r(r - i,/2 M ,[J_: r]! *(g"- r *) + • • • • (-i)y i, "- i,/2 «>(^ (3) 

(q n x - q n ~ 1 x)(q n ~ 1 x - q n ~-x) (q 2 x- qx)(qx - x) 

in which, as in previous papers, # [n] denotes (q n —l)/(q—\). 

In this expression the coefficients of the functions <fi(q n ~ r x) follow what may be 
termed the g-Binomial form, that is, they are identical with the coefficients in the 

(x+!/)(x + qy) .... (x + q"~ l y) = x n +[n\z"-hj + .... + W g *»-i>g"-y + , (4) 

By means of the operator A , the connection between the A -equations satisfied by 
certain g-series functions of x and the linear differential equations satisfied by the 
simple (q = 1) forms of such series is made clear, for from the equation 


a r/f./ \i <&(qx)-®(x) . q dx 

A{4>(a:)} = v * ' ±-' , we see x.k. — ± — 

qx — x q-\ 

also that with certain restrictions on the nature of <p(x) we may write 

i -»=x: + e- i >i ! s + <*- i >f 2 s + • • • ■ < 5 > 

and that xA is related to the differential operator x-p in a manner analogous to the 

relation between n and its g-form (q 11 - l)/(q - 1). 

§ 2. Basic Exponential and Transformation of an Arbitrary Power Series. 

The following simple example will suffice at this stage for illustration of the connec- 

tion between the ^-operator A and the differential operator -=- . 

Compare two equations 

I- W 

* Trans. Roy. Soc. Edin., vol. xli. pp. 1-28, 105-118, 399-408. 


On substituting 2,c,xc a+r m the first equation (a), we find, since (A^'" = [m]x m_1 ), an 
indicial equation [a] = 0, and a function 

X [a+l] + [a+l][a+2] + 

The equation [a] = 0, viz. (q a — l)/(<7 — l) = 0, has the doubly infinite system of roots 
« = , 2 7l . , (r = 0, 1, 2, 3, . . . .), (s = 0, 1, 2 . . . .)■ 

lOg g + 2S7T« 

From the first or principal root, which is zero, we deduce a solution 

Cf OP a* 

In connection with this function, which may also be expressed as an infinite 
product, we note that Boole's transformation of an arbitrary power series, viz. 

= e*{a + x. Aa + |p A 2 a + |- • A 3 a + J 

possesses a remarkably simple generalisation, in which E q (x) replaces e x , [2]! , 
[3]! , . . . . replace 2! , 3! , . . . . and g-operators replace the A operator of Boole's 

formula ; thus 

A p a* = (D-l)(D-g) . . . (B-q r - 1 )a n 
replaces the 

A\ = (D-1)\, Da„ = a n+1 , 

usual in Finite Differences. 

I think that these examples will suffice to show that for (^-functions the natural 
analogue of the theory of differential equations is the theory of the difference quotient 


A<I>(.r) = —sLJ. \-L or operator xA = ? = \xTf\ . . . (6) 

qx-x q-\ 

In the later portions of the paper several A-equations will be solved, which will 
show clearly that the theory of A-equations is exactly parallel to that of differential 
equations. Such solutions as Hankkl's solution of Bessel's equation have a place in 
the A-analogue of Bessel's equation. 

In previous papers the writer made use of an operator 

^ s a^){i&r*)\ • • • ■ I d&r*) i • ■ ■ • \W)\ — id®}- • • I, 

which could, however, only be applied to such special series as ^cx lr] , and was quite 
inapplicable to arbitrary functional forms. A comparison of the following equations 
will show the advantage which the operator A has over D ( ' l) . 

qsS*B»y - DM? + { 1 - [n] - [ - n - \}}xDy + [n][ - n - l]y = P ln ^(x) - P[„_ 2] (^) . (7) 


has a solution 

y ~ X [2][2n-lf X + ln]{) 

analogous to Legendre's series P„(a;). 

In the case of the A operator we have 

qx 2 A 2 y-^ 2 y+{l-[n]-[-n-\]}xAy + [n][-n-l]y = 0, . . (8) 

with a solution 

y - X [2J[2^T] + 

Of course, when we put q = 1, the right side of (7) vanishes, and both equations (7) and 
(8) become identical with Legendre's equation. The A-equation is, however, an exact 
parallel of Legendre's differential equation, and any transformation of Legendre's 
equation can be at once followed by similar transformations of (8). The presence of the 
two functions on the right side of (7) prevents our doing this in the case of (7). More- 
over, the A operation can be applied at once to product-functions, etc., which is quite 
impossible in the case of the operations denoted D ( '°. 

§ 3. Elementary Properties of the Operator A. 

So long as we confine ourselves to the direct operations denoted by A" [n a positive 
integer), it is clear that A as an operative symbol obeys the laws 

A m {A"$(a;)} = A"' + "{$(a;)} = A n {A'"$(a;)} 

A(m + v) = Am + Aw 

Aaw = aAw . 

Also in analogy with 

d , > dv , du , d ( u\ I da do | „ 

—(uv) = w— + v— and — - = { v— + u-— } -=- v 2 
dx dx dx dx\ v J I dx ax ) 

we have 

AU X V X = V qx &U x + Uz&Vz 

A ^ _ vAu x - u x Av x ^ 
v x v qx v x 

We note the following special cases : — 

A. a:" = [?<]x' !_1 , with analogies D.<" = nx n ~ x . 

A"E 9 (a.c) = a"E a (aa:) , D" Exp. (ax) = a n Exp. (ax). 

. , log a \ tm 1 

A log x = — 2_ji . _ ; JJ log x = — . 

q- 1 x x 

A(z- \)(x-q) . . . (x-q n - l ) = [n](;x-\)(x-q) . . . (x-q'" 2 ), B(x - 1)" = n(x - \f~\ 

A~E q (ax).x n = q n x n dE q (ax) + \n\x n ~^ q (ax) , De"V = ax"e ax + nx n ~ 1 e az . 

From which formulae corresponding to those of integration by parts may be found 

and (/-analogies of e n or n dx = T(n) formed, which I leave to a supplementary paper. 


§ 4. Interpretation of A - ". 

If in any function of x we substitute qx for x, let the operation be denoted by a 

a ... d 

prefixed symbol Q, thus Q = q x <ix in terms of the differential operator x-=- • 

so that 

Q-tf*> -/(£)■ 

Consider now the equation 

[VI ' 
4>(q"x)-[n]<P(q"- 1 x) + .... + ( - l)y ( '~" r i, [ iM '/'"'- 1 -) + • ■ • ■ (-l)Y"" ,_,, *(a;) 

A"*(a-) = -, - L ?, J-L M - ,,- J- . . (10) 

(q n x - (f x) (q 2 x - qx)((/x - x) 

In terms of the operating symbol Q, this may be written 

\Q n -[nW- 1 + q^k 1 ~-\ n - i - (-lJV*-*'!^) 

A' ! d>(V) = I LfJ: ' (in 

\ J ( 2 4»(»-«( 2 - 1)V " ' V ' 

The operating function 

| cy-MQ*-^ (-1)^-1,/. [»]i q»-'-+ (-i)v ( "- 1,/2 Q° | , 

may be expressed by the (/-Binomial theorem in the form 

(Q-1)(Q-<Z)(Q-? 2 ) (Q -'/'-'), 

since Q is an operative symbol obeying the laws 

Q*.Q\u = Q m+n .u=Q n .Q, m u, 

0,(11 + v) = Q.m + Q.y , 

therefore we write 

A « M _ (Q -i)(Q-tf)(Q-g 2 ) (Q -g-> ( 12) 

Now in the case of the " q " Binomial series, 

(I - a;)(l - qx)( 1 - q*x) . . . ( 1 - <f-hc) = 1 - [n]x + y^[" ~ ^x- - gsML" - ^I" " 2 ] x 3 + ( _ i f^n-m^ ( 

if we denote the product by (l - x)„, it is found that the equation 

(l-^-l-^ + jfcfcllLi- 

is valid for negative and non-integral values of n, provided (l — x) n be suitably 

interpreted from the following index relations, 

( 1 - *)„( 1 - q n x) m = (1 - x) n+m = ( 1 - *),„( 1 - q m x) n , 

(l-^'x),, - 1 - tf-arXl - r/-" +1 a;) ... (1 - 2 _! k)(l - jf 1 *) ' 

= 1 -[-»> + f/ " "J[ " " ~ ^a 2 - ad inf., 

which result may be applied in the case of the operative function (12) as follows 


§ 5. Ee verse Operation A -1 with (/-Finite Integration. 

The operative function 

(Q-i)(Q-?)(Q-? 2 ) (Q-'T 1 ) 

may be annulled by 


(i-Q- 1 )(i-aQ- 1 Xi-fl"Q- 1 ) ■ • ■ • (i-^Q- 1 )' 
which can be expressed in many different forms, one of which is 

<H i + (g^i) Q -i + (g"-i)(g" +1 - 1) I 

= Q-" + HQ-"- 1 + W[«+i]^^M[« + i][" + 2] Q [ 3^ 3 + , • • (13) 

so that the reverse operations indicated by the following equations will be legitimate, 
provided that the series obtained are finite or convergent. 

§ 6. Solution of A n w = (f>(x). 

A"w = $(a;), 
u = A~"<£(a;) , 
u={Q- + [n]Q- n - 1 + } {gM n -V(q - l)V*(a;)} , 

= 3 w«-*'(, i -l)»|Q-« + [„]Q-»-> + W^ + l]Q-»- 3 + . . .}.*"*(*). . . (14) 

Now, since &x m = [rri]x m ~' 1 , it is plain, that just as in indefinite integration, so here 
we may introduce an arbitrary periodical * constant of multiplicative periodicity at 
each performance of the operation A -1 . An example of such a function, which may 
be termed a function of multiplicative period q, is, 

... s ^ ( /27rlogx'\ , 7 ■ /27rloga:\ ) 

satisfying <f>(x) = <f>(qx) = $(.q n %) • 
We obtain then 

u = a + ai x+ .... +^_ 1 ^ + 2 w»- 1 '(2-ir{p*(^) + W^*(^r 1 )+ } . (15) 

The connection of these forms with integration is manifest, for on writing 1 + e for 
q, and putting n= 1, we obtain form (15), 

u = a + ai x+ . . . +«»-^- 1 + -{ (r ^ ) ^( f fJ + (T ^^( (1 ^y 2 )+ } 

and so retaining only terms involving e and neglecting e 2 , e 3 , 

u = a + a^x+ . . . . + a^xX 11 - 1 + L ^ \ex<f>(x - ex) + ex<j>(x - 2ex) + ]. 

It is easy to verify (15) in particular cases, for consider 

u=.- (a) then Au= — r (/3) 

(l-x) (\=x)(\-qx) 

* Cf. Boole, Finite Differences, pp. 46, 47. 

U A \(l-x)(l-qx)l 


If we now reverse the operator A in equation (/3), we obtain 



and (15) becomes 

,=tt,, "1f|H + #-:) + <-l4) + -'' 1 

On choosing the arbitrary constant c = 1 , and replacing the terms on the left by 

1 f 1 1 1 

series of powers of x. we find the coefficient of x r is [r~\ — { 1 + — + — +... '■ (q — 1 ), 

q' I q q rr J 

which ((/>1) is identically unity; so that a special value of the series is \+x + x 2 

+ . .... or 7— , , and the validity of the transformation is justified in this simple 
(1-x) J J L 

case. The result may be expressed also as 

(T~^x) = ( q ~ 1 H(l-x)(l -g_x) + (q - x)(q* -x) + (g 2 - x)(q3 -x) + ' ' ' I 

It will be noticed that such an equation as Aw = — cannot be solved in this form, 

though a solution is furnished by 

, A.(ff - 1) 

u = e+ —3 ' log x , 


reducing, when q = l, to c + Wogx. 

In the same manner, if we perform the operations denoted by A H on the function 

(1 - x)(l - qx)(l - q 2 x) .... {\-q m - l x) 

we obtain 

bPn = - [m][m+l] . ■ . . [m + n-l] 

(l-x)(l-qx) .... (1 - q m +"- 2 x)(l - q m+n - l x)' 

If we solve this by the inverse method, we find in general 
u = a + a x x + a 2 x 2 + . . . . + ^os" -1 + (q - 1 ) n qW>-v x n 

*s?[ n][n + 1] . . . [n + r- 1] • [m][m+ 1] . . . [m + rij-l] ,,g* 

one value of which is the particular expression for u given above. This value will be 
obtained by the following choice of arbitrary constants : 

a =l , 

_[m][?« + 1] .... [m + r-2][m + r-l] 


We notice that since A i»n.„_ Io §g . * 

and A n E g (ax) = a n E g (ax), 

the equation A m s/ + a x A™- V + + a m _Ay + a m y = <f>(x) 

may be solved by methods similar to those adopted in the case of differential equations 
of like form. It is obvious that by giving solutions of such equations as 

3>(A)u = *(x), 

in which <£ is so chosen that $ _1 (A) is capable of representation in such series as 
£, X . A", a great variety of q series and relations may be formed : for example, from 
such equations as 

(1 - 2 A)(1 - 2 *A) (1 - q^)u =f(x) , 

so that 

« = y r nir i i q n A n f(x) . 

£-[ro -l]![n]! tf J y ' 

§ 7. Forms of A n $( x ). 
In this section of the paper the following forms will be obtained : — 

A^(*) = 2g W "- 1, [W + r][W + r ~ 1] ' ' ' ' i [r+3][r + 2lr+1 W ^'(Q) > . . (16a) 

^ q ¥^H~ (rjl" W ' 

A "^)= i A CJMrxT 1+ + MsF + ■■■ \« x) * 

A r = i [n] n+r - [n][n - ll»+'+ ( - lVg*"'- 1 ' r ^ ! r -,, [n-sf+ r 

[ ' [n - s\\ [s\\ 

a. ( _ 1 \»/7i(»+l)(n-2)r«l I (^~ V 

+ •••■( ^2 L "Jfn + rigw*-!)' 

subject to the conditions that ^(x) and all its derivatives are finite and continuous. 
Theorem (16a) can be derived directly from the equation 

A nj < W J U- *(g"*)-[ttXg n - 1 «)+ (-lrgW'-^a!) 

1 v * /; ( 2 n a: - g"- 1 ^)^ 71 - 1 ^ - 2"- s a:) . . . . (qx-x) ' ' 

for, on replacing <$>(q n x), &(q n ~ 1 x), by Maclaurin series, we obtain from the 

numerator an expression 

|~3>(0) + 2 "a*'(0) + + ^W M (0) + | 

- [n] *(0) + g"-W(0) + +g__a; r $" , (0)+ j 

(_l)» 2 M«-i> _H! r#(0) + 3 "- s a!*'(0)+ +2!!LV*W(0)+ i 

[?t-sJ!|sJ!L /■'• J 

(-l)Y»(»-«("*(0)+ a*'(0) + +-y *"X°) - ]• 



Arranging this series according to powers of x, we find that the term containing x r is 

[2]! 1 +( l) [n-s]\[s)\ 

C'$" (0) j .„ _ r w -[„r|H-l) + [ n ][ n - 1L .(«-2|+1 + +(_iy M - „r(n-s)+ls(*-l) 

+ + ( - l)»gW«i-i) I . 

This expression is well known as being 

?! ^V-i)('/''- i -i)('/'- 2 -i) — (f-^-i) 

for all values of r > n. In case r < n or = w, the expression is identically zero. 

Hence, taking account of the denominator of (17), A n (p(x) may be written in the 

^ MIV-11 . . . . |>-ra + ll .... 

or, since [r][r — 1] . . . . [r — n+ 1] contains a zero factor for all values of r <.n, finally 
we write 

r=0 L J 

In the equation 


let us substitute 1 + e for q ; then, in case &(x + ex) is capable of expansion in the form 
of Taylor's series, 

A*(a;) = *'(*) + —*"(») + .... +-^-^"+ l \x)+ . ... 
2! ra+1! 

Similarly, from the equation 

it is easy to obtain 

(q 2 x - qx)(qx - x) 

A^(x) = (l± 2 -V'(.r)+|! , ( £ +2)(e+3)*'>)+ (19) 

w=» u— \,y,ll 

l(' +2 >"-w 

71 = 1 X ' 

or symbolically 

(g+iK ■ (g _ i)-i (g +i){( g +i)»-i }af-v <* Y +1 ran 

In the same manner, if we replace q by 1 4- e, in the general equation 

W ' (q n x - q n - l x)(q n - l x - q n ~ 2 x) . . . (qx-x) ' 


and suppose ($>{x) capable of expansion in the form of Taylor's series, we obtain 


= [*(*)+ + {(l+e)"-l}^ (r)(a;)+ J 

-w[^)+ + {(i+ e r i -i}v rw+ .... .J 

, Mrn-lir./ v , , {(l + e)"- 2 -l} r r . w/ * , 
+ 3 f2l! L^ ~^ ri L *y(*)+ ■ • • 

) +q in{n - 1] (q-\)"x n . 


+ ( _ l)» 2 W»-D^(a;) 

Collecting the terms in sets according to the orders of cj>', <p", , we find that all 

terms below the set containing </> (n) (x) vanish identically, since 

[n] m -[n][n-\] m + ql^^[n-2] m + + ( - lyv/''- 1 '"- >[n][l] m = 

for all values of m (an integer) < n. This identity is easily established if we consider 


A 2 <£(a;) = c<t>"(x) + terms in <£'" , </> (iv) , , 

then replacing q by 1 + e, and operating with A , we find 

A^(a;) = ciJ Li — ^AJ + similar terms in d)"', 


= c 2 <f>'"(x) + . . . . similar terms in <£ ,iv) , </> (v) , 

and by induction A n <p(x), when expressed in terms of <p(x) and its derivatives, contains 
neither <p(x) nor any of the derivatives below <p {n) (x), so that the coefficient of <p°' l) (x), 

2 -in(»-i)( 2 _l)-»r( 3 »_l)^_[ TC ]( 2 »-i_l)» + gWElzlJ( 3 »-2_i)'»- 

L L 2 J ! 

- ( - \fq^-^-\q - 1)»[»] ~\^ , 

is identically zero for all values of m < n (m, n being positive integers). 
The coefficient of ft n+r \x) is 

|~( g n _ i)-* _ [„] (g n-i _ 1)n+ r + q [n][n- l] (gB ., _ 1)B+r _ _ _ _ 


n + r\ 


Hence, taking account of the denominator in expression (21a), we find 

i *)=(^J tA '(ir tj ' 

d_\ n+2 






in which 

a = / r«i"+ r _ r«ir«. - 1 1«+* 4. . -i. i _ 1 \»«w»-«_ 


A, = | [«]"+'' _[„][« -!]»+>•+ + (-l)«gWM)JW [n-s] n+r + 

,(_ lW »- 1 »(»- a » W }_il=^_ )j . (25 ) 

+ i 

provided that the functions, on which the operations are performed, are capable of 
expansion in Maclaurin's form. The coefficient A r is the g'-equivalent of the number 
A".0" ( tabulated by De Morgan and Boole (Boole's Finite Differences, pp. 19, 20). 
The development of such expressions as A"$(a;)^a; may be shown to be dependent on 

A<>(.r)<%))={A^ + QA;}{A,«- 1 + QA>} .... {A. + QAV}!^)*^)} 
in which 

A q r4>(x) = <t>(qx)-q r 3>(x),A^ = \,,\n +1 . . . A^A, , Q and A 

operate on &(x) only. I leave the consideration of these forms, however, to another 

§ 8. Forms of Maclaurin's Series. 
In (24), putting 

A M *(a:) : 

we see that 


d\ n+1 


i A"*(a:) 1 = A $<"»(0) , 

= { W " MO -lf+ + ( - m^^^f.n - sf +...(- l)y»-*->] } J^,, 

= K ! $N(0), 

so that Maclaurin's series takes the following form in g-function theory : 

say *( a O = *(0) + ^A*(0) + ^ ! A**(0)+ 

and a conjugate form 

<D(a;) = $(0) + -^4.(A.0)0 + ^ ! <E>(A.0).0 2 + 

where <I>(A.0)0'" denotes what <I>(A)ce'" becomes when x = Q. 

The special development of this and the analogue of Taylor's theorem I defer to 
another paper. 


Part II. 

§ 1. Certain Infinite Products. 

In this section of the paper I propose to investigate the expansion, in series form, 
of such products as 

(l + UyC + a 2 x 2 +■'/.. + a„j: n V 1 + a x — + a 2 — +,...+ a,,-^ V 1 + a x —^ ■ 

+ aX 

toW 1 + W 

q J V 2 
(1 + 2qx cos + (fx 2 )(l + 2q 2 x cos 6 + g 4 x 2 ) 2 ( 1 + 2q*x cos 6 + q 6 x 2 ) 3 

( 1A )( 1+ ^( 1+ -)' 

( 1 + 2qy cos 6 + ij V)( 1 + V.? cos 6> + ? y ) 2 ( 1 + 2q 3 y cos + q 6 y 2 ) 3 ' 

= \ (x , y) + X^x , y) cos 2$ + \ 2 (x , y) cos 36 + 

The coefficients X , X x , . . . . X n (x , y) will be seen to have an intimate connection 
with the q generalisation of Bessel's Functions denoted by the symbols J w , ^ [n] , . . . 
in previous papers. 

It is easy to see that the A -equation 

A.u = {c + c } x + c 2 x 2 + .... +c n _ l x n - 1 }u . . . (27) 

is satisfied by the convergent infinite product 

U = n s(«l > a 2 ' a 3 » • • • ■ «» » X ) » 

= (l W+ « 2 * 2 + . . +«„*»)(! +^ + ^2+ .... + |;,»)(l + 3., + |2 x2+ ."..+*) , 

in which 

"l — C Q > 



"2 ^2~ C l' 

«„ = <i^>e 

n 1 ■ 

The general factor of the function is 

(1 4- a x xq~ r + a 2 x 2 q~' lr + . . . + a n x n q~ nr ) . 

§ 2. Expansion as an Infinite Series. 
An examination of the infinite product shows us that in case the product can be 
represented by a series of powers of x , such series will be of the form '^ i c l x r , 

= l + -?±2-x + c 2 x 2 + c l x s + 


In the equation 

n q (z)=-?- T b.- 1 \((i 1 + a 2 qx + a 3 qh?+ . . . + cg—V- 1 )^*) \ . . (28) 

q — 1 I / 

replacing FI 9 by the series ^c r x r we obtain 

^c,y= -V— A -1 \ (a^c + a 1 c ] x + a 1 c 2 x 2 + .... +a 1 r, m x m + ) 

+ (qa 2 c x + qa 2 c^x- + + qn 2 c m x m+l + . . . . ) 

+ ((] 2 a s c Q x 2 + q^CyX 2 -4 + q 2 a s c m x m+1 +...),. . (28a) 

+ {q n -'a n c^'- 1 + q n - l a n c^' + + q"- 1 a„c m ^ m+ "-')] ■ 

x m+ * 

Now A 1 (ax , ") = C + a— — =, (C being an arbitrary constant), therefore performing 
the operation A _1 on the terms of this series (28a), we obtain 

1 (ft \ x , ( l 

2 /, / \ -v3 

Z c ■-'''' = c + tt\ a ^ x + Hhv iCi + ga * G vw\ + q^\v lC2 + qa -'' + ' y " ffl3 '' o y[3]r + 

and we find, by equating the coefficients of powers of x, that 



!= ?^i( ¥l+2 r 

Cr = -^-^(a 1 c r _ 1 + qa 2 c r _ 2 + q 2 a ;i c r _ 3 + + q n ' l a n c r _ n ) . . (29) 

Subject to the convergence of the series, we write therefore 

• n >i>«2> a ?( , x) = l + ^^x + ^ 2 _^ _ 1} | a* + (q - IK } 

+ (g s _ 1 K ^ l)( g - 1) { a * + <g " 1 > a i a » + (g2 " 1)a ^ + {q ~ 1)(g2 ~ 1)as } 

+ (E7&r) { «i 4 + (?- l)a 1 2 o 2 + (2 2 - ] K 2 «2 + ('i" l)('/-l)¥3 + (? 3 -l)«i 2 « 2 

+ (2-l)(2 8 -l)a 2 2 + (ff 8 -l)(2 2 -lKa 1 + (3)(2)(l)a 4 } + . (30) 

The recurrence relation connecting n + 1 successive coefficients being 

Cr = ( »• _ i ) 1 a '' v - 1 + IWr-l + 9 2 <¥V- 3 + +2"~ 1 <W-»|'j • • ( 31 ) 

the series is absolutely convergent for all values of x, in case | q | > 1, for the series is 
seen to be the product of n absolutely convergent series formed as follows : — 

Resolve the polynomial (l+a 1 x + a 2 x 2 + . . . +a n x n ) into its factors (I + pfl) 
(1 + p 2 x){l + Pz x) ... (I + Pn x). 



n,(¥A • • M=n(i + ^)n(i + *£) & „( 1+ ^)' 

m=0\ Q /?n=0\ <± / m=0\ ({ / 

i 1 + PnW ■ pA 2 * 2 , I 

1 (q-i) v-i)(?-i) r 

and each of these is absolutely convergent for all values of x in case | q | > 1 , as well as 
for limited values of x in case | q | < 1 . 

§ 3. Product of n, Theta-Functions. 

In this section I propose to obtain a theorem, which may be stated as 

6(a! , x)Q(a 2 , x)Q(a 3 , x) Q(a n , x) = A + Aj cos 2x + A 2 cos 4a; + 

The functions (a 1 x)Q(a 2 , x) each reduce to Jacobi's function 0, in case 

the parameters a v a 2 , ct 3 , are each equal to unity. 

We have above obtained the expansion of the infinite product n <? ((x 1) a 2 , a 3 , . . . a n , x), 
in the form of a power series ^c r a: r , viz. — 

q m q 2m q Zm q nm J ((/ - \) (cf ~ l)(q 

Change q into q' 2 , and let (l + p 1 x)( 1 + p 2 x) . . . (1 +p n x) = (1 + a 1 x + a 2 x 2 + . . . 
+ a n x n ). We write then 

?n=oo m=oo m=oo 

U q -2{a,a 2 a z a n x) = n (1 + Pl q* m x). II (\+p 2 q im x) n (l+p^x), 

m=0 m=0 m=0 

J\ q -'i(a x , a 2 , a 3 a B , a;<).II fl -2(a 1 , a 2 , a 3 a rt ,^" 7 )= II (1 +p 1 a»/ am (t + ^~ 1 ) 

+ l 9 1 2 a;Y m ) n(l+p n a^ 2m (^ + r 1 ) + Pn 2 a;Y m ). (32) 


Now p v p 2 , p n are independent constants, for we may choose n arbitrary 


(1 +p 1 x)(l +p 2 x ) • • • (1+/V) to form a polynomial (l+a 1 x + a 2 x 2 + . . . +a n x n ). 

Replacing H q -z(xt) and H q -2(xt~ 1 ) by their series expressions ^?c r x r tf, and '^c r x r t" > \ we 
obtain from the product of these series 

n 8 -2(a 1 a 2 a 3 . . . a n , a*)II ? -i!(a 1 a 2 . . . a„, a# _1 ) = J + 2J 1 (< + r 1 ) + 2 J 2 (i! 2 + f~ 2 ) + (33) 

J . J l5 J 2 , , are infinite series functions of x, reducing in special cases to 

Bessel's functions, 

J ° - ' + ggjy + {<t - dv - i/ ,jV + "* - ™ + <34> 



The coefficient of x 2 '" is c m 2 , which is determined by the recurrence relation 


"2 a -i C m-2 T" —£ a 3 e m-3 +•••• + 

1 ) 

2^T 2 a n c m-n j ' 

. (35) 

J m = x m {c m+<) c + c m+1 cx l i + c m+2 c Q x i + }, . . . (36) 

in which the coefficient of x'" +2r is c m+2 ,..c. 2r , in which c lll+2r and c 2l . respectively are 
determined by the above recurrence relation (35). 

§ 4. Expression for O n (x). 
We have now established the form 

U \ 1 + 2q im x Pl cos 20 + q im x> Pl 2 i S j 1 + 2g 2m *p 2 cos 20 + <f n a; 2 p 2 2 1 

m=0 I J 7)1=0 I J 

n •! 1 + 22 2m zp n cos 20 + 2 4m x 2 Pr! 2 !• , = J + 2 J x . cos 26 + 2 J 2 . cos 40 + . 

m-0 ( J 

• (37) 

the constants p 1 p 2 . • • /o„, being connected with the constants a l9 a 2 . . . . a„. para- 
meters of the J functions, by the relation (1 +/) I a;)(l + p 2 x) . . . ( 1 + p n x) = 1 + a x x 

+ a 2 x A + . . . +a n x n . 

Thus, if we choose p 1 = p 2 

1 , a 1 = n, a 

_n(n— 1) 


put g~ l for x, we shall obtain the n th power of Jacobi's © function 
@ V2K0\ I n(l- 2 *»)iTj + 2J 1 cos25 + 2J 2 cos4e+ . . 

in which 

a n = 1 , and 

• (38) 

• (39) 

j _ , , ra 2 ra 2 {w 2 g' 2 + rag 2 + ra 2 - ra} 2 + c r 2 

° - <z 2 (<z 2 -i) 2 5 4 (? 4 -i)% 2 -i) 2 7 r • • ■ . , • 

the other coefficients being determined by means of the recurrence relation 

J»' = g-'»{c m c + ( ? - 2 c m+1 c 1 + g- 4 c m+2 c 2 + } (40) 

The simplest case of the product II^o^ . . . a n , x) is when a x = a(g — l)/q, and a 2 a 3 , 
. . . . , are each zero. In this case the above product and series become 

which we saw in (9) was a solution of Ay = ay, and was there denoted E 5 (asc). This 
product and series is, of course, well known as 

( 1+ *)H) 

1 + 

(2-1) (q*-l)(q-l) 

If in the series E g (ax) we replace g by g -1 , which we term inverting the base, we find 

ax . a'x 


E q -i(az) = l + ™+q™+q 


x a x 

[1]'*[2]! * [3]! 


with for limited values of x a common region of convergence with E 7 (acc). We shall 
subsequently require the properties 

■E q (ax)E q -i(-ax) = l (41) 

W Vfl«) - 1 + ffiW <" + 'X« + *% * + 

§ 5. "Double Product Theta-Functions." 
Consider a function ^(x), of the form 

<t>(x).<t>(*).<t>(?-S ad inf. 

+W-%}+($ ad inf - 

In case the infinite product is convergent, we may write 

**>-$$$&<>»• ■ ■ ■ ■ <«> 

and by means of (15) 

*(*) - e + «*] ~ *«> «(«) + *W ; y ^ + (44) 

xf;(x) \qj Jx\ \qV 

Let us now consider the special function &(x), in which <p(x) = ~R q (ax),y\/(x) = 'E, q (bx), 
so that 

•« -!©•«■ • • •_..■'• ■ <45) 

From (45) we have at once by means of (41) 

*/ \ / i , ( a - b) , (a-b)(a-qb) » „ , , (a-b) . . . (a-q'~ l b) ,. .. ) _, . .... 

*(</») = \ !+ V iyj -^+- r 2 \, V -' c2 + ■ • • • + V '- P^p = >-q'x + ■ ■ • [*(«), (46) 

so that, in case &(x) can be expressed as a series of positive powers of x, as seems 
possible (a priori) from the nature of the infinite product, which is of the form 

\ + ax\(\+-xX(l + ^rx) 

'S tlS tl , .... (47) 

i + ^)(i + ^)(i + !*) 3 

we shall have the following relations to determine the coefficients in the expansion 

<&(x) — c + L\X + c 2 x 2 + , 

( 'o = l . 

Cl i 7TT = tvz ' 

, a-b Act- b)(a - qb) 9 ,, 

c ' 2 TiT^ 1 — [W~ = ' 

,. (a n _ n _ (a - &)-. . (a - ft) (« ~ ff& t 2 ,. , . (« - &)(« - qh) . . ■ (a- g-^) - , . 



which determines the form of the series expansion of the infinite product 

I + &% " 1)V ! + M<1 - })\y x + te(g-l) > * 

or, as it may be expressed in terms of the function E 9 (.x), 

E,K)E,(f )E,(- 

ad inf (50) 

W E,( 4 |)E,g) 

§ 6. Double Product Theta-Functions. 

Some interesting relations exist between the coefficients c^c^ and the 

generalisation of Bessel's Function denoted J [n] (^) in previous papers (loc. cit.). 
From the relation 

^-JS'^) < 51) 


we form 

', axt\f, axt\ 2 f, , aa:f\ 3 

, g /V i-l^ £-^ = 1 + c.xt + c<,xH 2 + 

1+ ^)( 1+ -)' (l+ -y 

q A + f )\ + ^f) ,_,._ 

fcrf-Vt . bzt-W, bxt- 1 ^ -l + ejZt 1 + c. 2 xH 2 + 


(1+ ^ )(1+ ^ nl+ 

in which 


(1 + 2axq- i cos 20 + a 2 x 2 g~ 2 )(l + 2axg~ 2 cos 20 + a 2 x 2 y- 4 ) 2 ( 1 + 2aa<T 3 cos 20 + a 2 x 2 q- 6 ) 3 

(1 + 2 toy" 1 cos 2(9 + l> 2 x 2 q- 2 )(l + 2bxg~* cos 20 + 6V 2 y" 4 )' 2 ( 1 + 2bxq~ 3 cos 20 + bh?q- f >) 3 

= 2<VZ r e 2r * 9 . x 2,^'e""" 9 = A o (:c) + 2A. 1 (x)cos20 + 2A. 2 (a;)cos40 + (53) 

or, by changing q l into q 2 , 

/ 1 • + 2ayry 2 cos 20 + aVt/ N VI + 2a.«/> cos 20 + a 2 / 2 / \ 2 /l + 2axg 6 cos 20 + a 2 xV 2 \ 3 ad jnf 

V 1 + 26ay/ 2 cos 20 + b 2 x 2 q* ') VI + 2fov/ 4 cos 20 + bWq* ) \1 + 21mf cos 20 + b 2 x 2 q™ ) 

= /a o (a;) + 2/^(a:)cos20 + 2/* 2 (a:)cos40+ . . • (54) 



in which 

/jl q (x) = c 2 + Cj 2 x 2 + c?x A + + c r 2 x' r + . . . 

pjp) = c n c x n + c n+i c i x n+2 + + Cu+rCrX n+2r + 

c r being determined by the relation 

ft, r (I-? 2 )'- 1 (l-r/)(l -^)'-' i (1 

;i-a a )(i-2*)(i-s e ) 

c,_, + 

+ (a- b)(q*a-b) ... (g 2 '- 2 a-b ) 

(l-2 s )(l-2 4 ) • ■ (1-2) 

2F~ '0' 

§ 7. Coefficients /«„, ^ related to X n . 
Putting 6 = 0, we obtain 

(1 4- 2axcf cos 28 + a^V) 1 ^ 4- 2axq i cos 28 + aWq s ) 2 
v (x) + 2v 1 (x) cos 28 + 2v 2 (x) cos 40 + 

in which 

v„(x) = c n c x n + c n+1 c 1 x n+2 + 

l\-q lr ) ac r _, , , a 2 c r 

«* (1-2 2 "(1 -g 2 )(l -g 4 ) 

Similarly, we may show that 

+ q r 

a'c n 

(l-g 2 )(l-g 4 ) (l-2 2 


( 1 + 2axg 2 cos 26 + u*sfyf(l + 2axq i cos 2d + a 2 x 2 q s ?^- + ^axq 6 cos 26 + aV^ 18 ) 3 
= A o (a;) + 2A 1 (a;)cos204-2A. 2 (a;)cos404- , 

in which 

Xjx) = C n C<p n + c n _ l c l x n+ ' 2 + 

a 2r r ((fi-iyitfl- l)(^-l)7 („2-\)(ai-i) .. 

(g 2 -l)(g 4 -l) • • • (2 2r "l) 





T ( 2 2_i) (gt-l)(2*-l) 

It is not difficult to see that 

"ofc) = \(*)/*o(«) + 2X l(-«>l(-« ; ) + ^fcW*) + » 


"rfc) = 2, 0*r+l(»)M«) + /V(*)^r+l(*)). 

We shall now obtain certain relations connecting these coefficients with the q 
generalisation of Bessel's Function. 

§ 8. Connection of Coefficients with ^-Bessel Coefficients. 

To show the connection between the coefficients A., n, u, and the q generalisations 
of Bessel's coefficients we proceed as follows : — 


\ l _ ax(q-l) ' 
2 - 

E>x) Ei («*)e,(^) 


j + ax(q - I) 


*(g - i) T 
<f J 

"l + Kg -1)1 _ ! + Mg-i) ! 2 ^ + Mg - 1) 1 

g JL g 2 J l g 3 J 




We see that 

/ x \ (l + a X q-i)(l + au:g-*)*(l+axq-sy . . . . . (&Q) 

"\q-\J (\+b.rq-%l+bxq-y(l+bxq- s ) s ' V ' 

and that 

* ( ? v V (* + aM f)( l + azq^a + a*q 6 ) 3 

*""Vl-flV (1 - bjq*)(l + krq*)\l + bxq^f ' 


. ( ,,,■ v E ! ^f$M5 ! ) 


? y . . . (61) 

r 2 ' 


The expression on the left side of this equation is by (47) and (49) 

[1 +axq*(t* + t*) + q»aVl 1 [ l+aa»j 4 (< 8 + f 8 ) + o^g 8 ] 8 (62) 

= / , ( a: ) + / x 1 (^ 2 + r 2 ) + / x 2 ( a; )(^ + r'')+ 

The expression on the right side of equation (61) may be given in various forms, owing 
to the following properties of the function E g (x), 

E»xE rl (-^)=l, (63) 

W_ E / a v xE (Ml, ("-&)| ( fl -fr)( a -g 6 ) + . 

We shall first transform the expression on the right side of (61) into 

V/-I/ Vg 2 -1 / , . . . (64) 

Now I have shown * that 

= J [ o,(A) + (^ + nJ [1 ,(A)+ +(f m + t-*»)J w (X)+ , 

Jt " ](X) = f§ {2w + 2r}!{2r} ! ' 

Replacing in (64) -^ : by .=-= , for convenience of writing, and -^- by ..— = , 
we^have finally 

/ , (; C )+ / x 1 (*)(* 2 + n+ _J| 0) (^) + (< 2 + nJw(^)+ (65) 

*>(-) + /*i(-)(< 2 + '- 2 ) J ^) + (t 2 + ^)J «+ 

* /Vw. L.M.&., series 2, vol. ii. p. 196. 




_ , (m - v)x (u- v)(u - q 2 v)x 2 (u-v)(u- q*v)(u -q*v)x? m) 

" 1 + -[2T + [2p] "M4][6] 

so that the expression (64) may be written 

1 | (u -»)■>•? + (u-v)(u-q*v) r2fi+ 1 y Y x , (u - v)xt- i + ( y - «)( tt - g 8 ")^- 4 + 

[2] [2]L4] 

[2J [2][4] 


= | + (< a + «- s )li + (^ + r ')-la+ 

in which the coefficients 3f are ? analogues of Bessel's function J n (u— v). We see that 

„ (tt - t-)(tt - g'tQ ...(«- 7 g "-^ ) / , ^ (» - »)(« ~ g 2 "") ^ (» - «)(« - g 2 »)(" - g^Xt* - q in+ 'v) , I , fi8 v 

|n [2][4]...[2»] I [2n + 2][2] [2n + 2][2n + 4][2][4] J' 1 °' 

The q, or quasi-addition theorem, for this function is, in the notation of a former 

Io(« , + »)= J ra (»)I m («) + ffiMhfc) + £fW«)ImW + 

in which 

J[ " ,)(m) = 2{2« + 2»-}!{2r}! = 2 I> + r + 1)I>+ l)(2) n+r (2) r ' 
^ lm]W ~ 2-{2m + 2r}!{2»-}! 2 

■ 3\ 


3f[„] is not a function distinct from J [n] , but can be derived from J [n] by inversion of the 
base q. It will be more convenient in subsequent work to denote these functions J q (n , x) 
and J q _ 1 (n , x) respectively, which will show explicitly that they are g-functions, and in 
what manner they are related to one another. For a detailed discussion of their 
properties I refer to Proc. L. M.S., series 2, vol. ii. pp. 193-220, vol. iii. pp. 1-24; 
Trans. R.S.E., vol. xli. pp. 401-408, pp. 105-118; Proc. R.S., vol. lxxvi. A., pp. 
127-145, vol. lxxiv. pp. 64-72. 

From (61), (64), and (67) we obtain 

H Q (x) + (t* + r i ) H . 1 (x) + (t* + t- i )^(x)+ 

Mo(|)+(^+nM 1 (|)+(^+^-v 2 (|)+ 

= $ (u,v) + (t* + t->)Uu,v)+ , . (69) 

whence we derive 

/*o(*)=/*o(-f}&> + 2 /*i(f)li + 2 ^(|)l 2 + 

ft^HZ {^(|)l» + r + /w(|)l,} (70) 


Similarly, from (55)-(56a) we obtain 

y [y) = v o( x ) J m{ vx ) + 2v i( x ) J d vx ) + 2 M x ) J [d vx ) + » • • ( 71 ) 

"»( ?) = 2 | "'■(?) J [«+r]( v ») + "»+r(a')JM(^) [ , 

K( x )=Ky}) J U i ' x ) + 2k i(jj J m( ux )+ 

K( X ) = 2, { X 'Q) J Cn+-'j( M: '') + ^wQJmC"*) > , 

in which 

ag 2 hq 1 

1 _ ? 2 ' ! _ ? 2 

Part III. 

Certain a -Equations and their Limiting Forms as Differential Equations. 

§ 1. The General Hypergeometric Series. 

It will only be possible in this section of the paper to state the forms of a few 
types of A -equations and their solutions, without entering into very detailed analysis. 

<l>[xA.y]- ±i/,[xAy] = (72) 

has an infinite number of solutions of the form 

y = A { *>+ *M r r"+i + *Wf« +1L «.+2 + I, . . (73) 

in which 

<£[a:.A t y] denotes Ctf/ + G 1 z.&y + C 2 x 2 A' 2 y + .... + C> 5 A 5 ?/, 
i/{a;.Ay] „ C' y + 0\xAy + C> 2 A 2 s/ + .... +C t x t A'y, 

for, on operating with (p\_x. A 2/] on a series Vc r aj 0+ '', 
we obtain 

2« a+r «r{C + C 1 [a + r] + C 2 [a + r][a + r-l]+ . . . + C,[a + r] . . [a + r - s + 1] } = ^c^+^a + r] . 

Similarly operating with - \|/-[sc. A?/] we obtain ^c r x a+r ~ 1 \|/[a + ?'~|, so that the 

solutions are furnished by the function (73). in which a is any root of the indicial 


C' + C\[a+r] + V'„[a-{-r][a+r- 1]+ . . . +C',[a + r] . . . [a+r-t+l]. 

The general Hypergeometric Series of ^-form is a special case of this equation and 
depends on the following theorem, which may be proved by induction : — 

« ««•*-] -«• (^X^ 1 ) • • • • (^> 

then II„[a + m] = A + A 1 [m] + A 2 [m][m- 1] + .... + A„[m][m-1] . . . [m-n+1], 


in which the coefficients A , A x , . . . . are independent of m and are given by 

For example, 

[a 1 + 3][a 2 + 3] [a„ + 3] _ [a, + 2] [a, + 2] U + \] .... [a„+l] _ [qj .... [q w ] 

[3]! [2]![1]! ~ +/ [1]![2]! Z " [3]! ~' 

the ^-analogue of a well-known algebraical transformation, 
if the A difference equation 

r=n r=m 

in which A r is defined above, 

and B,.= g(-1)^ « [? ._ s]![s]! , 
be compared with form (73) we see that on substituting ^c,x" txr , 

<f>[m] = A + A 1 [m] + A 2 [m][m- 1]+ . . . +A„[m]. . . [m-n+1], 
= n„[a 4- iii\ , (as defined above) . 

In the same way 

i//[m] = B [m] + B 1 [?»][?n - 1] + . . . + B„[?ra][m- 1] . , . [m-raj, 
= [m]n n [f3 + m-l], 

so that the solutions are given by a function 

n n [a + m] +1 n w [a + OT ]n„[ a+w+l ] m+i n > 

[m+ljn„,(j8 + ni] [m+l][m + 2]n„ 1 [ J 8 + m]n„ 1 [ j 8 + 7» + l] ' ' ^ ; 

in which m is any root of 

[m]n„[/3 + m-l] = 0. 

The principal roots are 0, I— /5 1} 1— /3 2 > 1— £ 3 , , and corresponding to each 

of these n l principal roots there is doubly infinite system as in general 

m=l-& + , grT/ nj , (r,*=0,l,2 .... ad inf.) (s = 0, 1,2,3 . . . ».). 

lOg (/ + 2^7rt A/ 

The series from the principal root zero is 

, /= i + hi M • • • [ a «] , + [ a i][ a i + 1 l • • • K][q»+i] o 7r , 

7 [iMLAJ • • W WMA+i] • • • L&I&+1] ( } 

From the principal root 1 - & we shall have 

The particular case 

(^-l)( t /-l) 

' 1+ (q-l)fr-l) X+ 

is a solution of 

[a][p]y+{q«[p+l]x + qP[a]x-[y]}Ay+{q*+P+h*-qyx}*y = 0. . . (78) 


§ 2. Special ^-Difference Equations. 

Three linear differential equations whose solutions enter into many interesting 
relations with one another are 

x*y" + xy' + (x 2 -'n*)y = 0, . . . , . (79) 

vty"" + 6.<V " + (7 - 2m 2 - 2n 2 )xV + ( 1 - 2m 2 - 2n 2 )xy l + (x 2 - (m 2 - n 2 f)y = , . (80) 

xY" + 6*V" + (7 - 2??i 2 - 2n 2 + 4x 2 )xV ' + ( 1 - 2m 2 - 2w 2 )x// 1 - (to 2 - rc 2 ) 2 y = . . (81) 

By means of (74) it is not difficult to construct the corresponding A-equations : 
among the solutions of — 

(79) is Bessel's Function J„(x) , 

of (80) is J,„ . n (x) = £ ( - I ) 2 r(?n +n + r+ ^ T —— i - i)T{n + r+l)T{j . + l )2 ,„ +M+2 ,. , 

f«m TM WWV T{m + n + 2r+\) x m+n+tr 

ox (M) ' ) >" { - t > x ' ) "\ x >- 2sT(m+7i + r+l)T(m + r + l)T(n + r+\) T(r + l)2" , + n+ip ' 

The analogous A -equations are 

r/x 2 A 2 ?/ + (l -[«]-[- w])xAy + (x 2 -[ra][ -«])y = 0, . . . (82) 

7 G x 4 A 4 // + r/([l] + [2] + [3] + [to + n] + [to - re] + [re - m] + [ - m - n])x s A?y 

+ cx 2 A 2 y + bxAy + x 2 y + ay = 0, (83) 

</ 6 x 4 A 4 */ + <f([l] + [2] + [3] + [m + n] + [to - re] + [re - m] + [ - m - n].c s A 3 y 

+ (ex 2 + [2] 2 x 4 )A 2 y + bx\y + ay = , (84) 

in which a = \_m + n][?n - n][n - m][ - m - re] , 

b — [to + 7i+ 1][to -7i + l][re - to + 1][ - m - n + 1] - [to + ?j][to - re][re - to][ - w - »] , 

[?;* + re + 2] [ - m - ?2 + 21 r ,-, r , n , T?" + «1 • • • • r - m - rel 

c = L J „T^ J-[m + n+l] .... [-TO-n+l] + f/ J rg-., L — ', 

which, when g , = l, reduce to the differential equations (79), (80), (81) respectively. 
We solve these A-equations by substituting a series ^c s x a+2s and, in the case of 
equation (82), obtain an indicial equation 

[a + n][a - re] = , 
and an indicial function 

y =6 '° { *" ' [a + re+2][a-re + 2] + } ' 

The principal roots of the indicial equation are +n , — n. From these we obtain two 
principal solutions, which are not distinct when n = ; also in case n is an integer, one 
of the solutions becomes formally infinite. By a suitable choice of the arbitrary 
constant we write down the solution corresponding to the root n, as 

'-SK-' V+^wi "^^ <85) 

{2?*}! = [2][4][6] . . . [2r] in case r is a positive integer. The form of {2n}\ 
when n is not integral, I have discussed in Proc. L.M.S., ser. 2, vol. ii. p. 195. The 
cases mentioned above, when one solution becomes formally infinite, or when the two 
solutions are not distinct, may be treated in the manner usual in finding solutions for 
differentia] equations in similar cases. The following is an example : — 


§ 3. Solution corresponding to Hankel's Y n . 
Such a solution of equation (82) is 

y = C 2 ( ~ * ) '[2][4] . . . [2r] . [2 - 2»] . . . [2r - 2ref 2 '" 


+ D log x"V ( - l) r i~ -,,'tn , a 1 

o ^jV / 1 2?-}! {2ra + 2r} 


"^"^ 1[2] + [4] + • • • [2r] + [2ra + 2j + [2 M + 4] ' ' ' [2n + 2r]f {2r}\ \2n+ 2r\ 

The constants C and D not being independent, I do not propose to discuss such 
solutions of the A -equations, but merely indicate that there are such solutions, as 
showing the parallel nature of the theory of these equations with the theory of ordinary 
differential equations. It is clear that Hankel's solution of Bessel's equation can be 
derived from the above in the limiting case 3=1. 
The solution for n = , is 

^ I 2<f 2<f 2q 2 ' \ .<■"• 

y = cJ q (0,qx)logz-c^ \ [2j + [iJ + ■ ■ ■ ■ +[27] / [2]»|4]» . . . [2sf 

I do not propose to discuss here the solutions obtained from the other roots of the 
indicial equation, namely those corresponding to 

a=±n+ 2 ™ . ,(r = 0, 1,2,3, ....). 
. . — — log q + 2<tti 

§ 4. Solution of ^-Difference Equation satisfied by J m . J n . 

In the case of equation (80), on substituting ^c s x a+2s for y, we obtain an indicial 

[m + n + a][m -n + a~\[m + n - a][ - m - v + a] = , 
and an indicial function 

J ° I [a + m + n + 2][a--m + n + 2][a + m-ii + 2][a-m-n+ 2] I ' 

and from the four principal roots of the indicial equation we derive four principal 
solutions, which, however, are not distinct in case m and n are both zero, and which 
become formally infinite in certain cases, so that special solutions involving log x would 
be necessary to replace these indeterminate forms. The four principal solutions, when 
distinct, will be denoted J q (m , n , x) , J q ( — m , n , x) , 3 q (m , — n,x),J q ( — m, —n,x), • 

J«(".".«) = Z(-D , {2BI + 2yt + 2r}!{2w + 2 7J T { 2 )> + 2r}lW l- ■ (87) 

As before {2r}! denotes [2][4] . . . [2r]. 

This function possesses many interesting properties, which will be given in § (4) 
et seq. It can be derived from the product of two functions J q (n , x). 



In the case of equation (84), if we substitute ^c s x a+2 * for y, we obtain an expression 

2, J [a + m +n + 2s][a -m-n + 2s][a + m-n + 2s][a - m + n + 2s]c s x a+2 » - [2a + 4s][2a + 4s - 2 ]<yc a + 2s + 2 I , 

so that the indicial equation is 

[a + m + n][a -m- n][a + m - ?i][a -m + n] = 0, 
and the indicial function 

/■-./«.+ [2a. + 4][2a + 2] ) 

•^ °l [o + m + » + 2][a-m-n+2][a + m-ra+2][a-w+re+2] J 

Choosing c n = -7 ^r-p — ,- , we obtain from the root m + n of the indicial equation 

5 ° {2mj! {2w}! ' H 

a principal solution 

^ {2m + 2n + ±r}\x'"+ n+ * r , g 

V ^>{2m + 2n + 2r}\[2m + 2r}l{2n + 2r\\{2r}\' ' K ' 

which I have shown"" to be the product of two q generalised Bess el functions, so we 

y 1 = eJ q (m,x).J q _ 1 (n,x). ..... (89) 

The meaning of J 9 _ x will require explanation. We have in (85) 

X m [ X* xt ) 

Jq ( m '^ = {2l^}\ I 1 ~ [2m + 2][2] + [2m + 2][2m + 4][2][4] " j ' ' 

If in this series we replace q by q-i a new series is formed, viz. — 

/ _ 1 V___f n „a+-2,im+r) 

K ; {2m + 2r}!{2r}r 

which I denote Jg.^m , as). Both J g _! and J 9 have a common region of convergence. 
In previous papers I have used the symbols J [m] (as), q m2 3m( x )> to denote these series 
respectively, but I think it is better to use the notation J q (m , as), as being both easier 
to print, and as showing explicitly the nature of the base. Moreover, it is apparent 
that the functions J q , J g _i are not distinct functions, but may be derived immediately, 
the one from the other, by inversion of the base q. 

xi nt / \ mw iv {2m + 2n + 4r\\x m+n+ » r /Q1 , 

J.(n,s).J 9 -,KsH g Z(- l ) { 2 m + 2n + 2 r}n2m + 2r}l{2n + 2r}l{2 7}l- ' ' (91) 

t/ \t / \ m»V/ i\r \2m + 2n + ir}lx" l+n+ ' 2r /Q0 . 

J q (m , x)J q - l{ n , x) = <t 2( " 1 )' {2m + 2 , + 2, }!{2OT + 2 ; }!{2M + 2 , }!{2r} ! ■ < 92 > 

It is interesting to notice that 

J v (n , x)J q -i(m , x) 
J q (m , x)J q -\{n , x) 

* Proc. Roy. Soc, vol. lxxiv. p. 67. t Ibid., pp. 67-68. 



is independent of x and is equal to q n2 **, from which many curious results may be 
derived. We pass on, however, to give the four principal solutions of (84) — 

>J\ = ?"%(« , x)J q -i(m , x = (fZ q {m , x)J q -i(n , x) , 

Vi = 2 m2 J«(rc . ar) J 9 -i( -m,x) = (f 2 J q ( - m , x)J q -i(n , x) , 

V 3 = (Z'" 2 J 9 ( - n . *)Jj-i( -m,x) = q n2 i q ( - m , «) J g -i( -n,x), 

y 4 = q r " 2 J tJ ( - n , x)J q -i(m , x) = q n \\ q {m , x)J q -i( -n,x), 

which may be expressed in a great variety of forms, as follows. 
It is easy to establish a theorem for q functions analogous to 


t / x # n I ■> , • , (2» + 3) , , , (2n + 5) ., , , ) 

>(2« + 2) 


which is a well-known result in the theory of Besskl's Function. 

The corresponding result for an arbitrary power series, and the q- generalisation of 
the same, I propose to discuss in the supplementary paper. 

The (/-analogues of (93a) are 


E q (ix)J q - 1 (n,x) = q*^{ 

[2][2n + 2] [3]![2n + 2] 

1 + KB + ; 



J q -i(w , x) = , 8 E g -i(Kc) _ if 

J q (n , x) \\{ix) E q (ix)E q ( - ix) 




and so the solution y x can be expressed in the form 

y _ lf c>+n ^ v {n,x)3 q (m,x) 
Jl + E q (ix)E q (-ix) ' 

so that if we denote this solution of (84) by <J> 9 (m, n, x), we have the following forms — 

%{m,n,x) = q'" 2 J q (u,x)J q _ 1 (m,x), ....... (97) 

= q nll J q (m,x)J q _ 1 (n,x), 

= q"> ,+ " 2 J q (m , x)J q (?i ,x)±{ E q (ix)E q ( - ix) } , . 
= qi^^{J q (m , x)J q (n,x)J q ^(m, x)J ^(n , x)} ! > , 

{2m + 2n + 4:r}\x m+n+2r 

= 'Z m2+ " 2 Z(- 1 ) r T2 

{2m + 2n + 2r}\ {2m + 2r}\ {2n + 2r}\ {2r}\ : 





= 2+nJ fy ( _ i y H'» + 4r}!^-"^ P ly, , , {4 W + 4r}!^^ ) >> , 

)^- V ; {4m+2r}!{2m + 2r}!2{2r}! J I ^ V ' {4rc + 2r}\ {2n+ 2r}! 2 {2r}! J ' V ; 

§ 5. Relations between various Series. 

Several interesting q series may be deduced from these relations. By equating 
(97) and (98) we obtain 

'/" 2+ "'{J,(»M| x)J v _,(m, aOJ^n.aOJf.ttfi, *)} 

= gww ly/, v, {2m + 2n + 4r}\x'"+ n+ir ) 

* \^ y ' J2»n + 2* + 2r}!{2»t + 2»-}!J2n + 2r}!{2r}!j 



which by (87) may be written 

|y ( _l V {4m + 4r}!^-"+ 2 >- l|V(-ir < 4w+ ^V-^L \ /m 

I ^ V '{im + 2r)\{2m + 2r\\{2m+2r}\{2r}\\\^ y ' {4n + 2r)\ {2n + 2r}\ {2n + 2r}\ {2r}\ J VP; 

= I y C - 1 V {2m + 2n + ir}\ x m+n+2r ) J 

~ I ^ { { 2m + 2n + 2r} ! { 2n+2r}\ {2m +2r}\ {2r} ! J ' 

whence, equating coefficients of x 2m+2n+2 '' , 

Z{4wi + 4»-}!{4n + 4i/-4?-}! 
(- 1 )' {4»i + 2r}!{4rc + 2v-2r}!{2m + 2r}! 2 {2n + 2v-2r}! 2 {2v- 2r}! {2r}! 

v {2m + 2w + 4v-4r}! {2m + 2rc + 4r}! 

= ^' ~ l '2m + 2n + 2v- 2r}\ {2m + 2n + 2r}\ \2n + 2v- 2r\l {2n + 2r)\ 

{2m + 2v-2r}!{2m + 2r}!{2v + 2r}!{2r}! 

in which {2?-}! = [2][4] . . [2rJ and [>■] = (q 2r - \)/(q - 1). This identity holds when 
m and n are not integral, if \2n] ! be interpreted generally by a Basic-gamma Function. 
When q=l, the series reduce to special hypergeometric series (x=l), with eight ele- 
ments in the denominator and two in the numerator of each term. It would be tedious 
to obtain further identities. I need mention only, that by considering such products as 

J q (?i , xt)J q (n , ■jrt-')J q _ l ((v , xt)3 q _i(v , xt~ l ) 

a great variety of relations can be obtained.* 

§ 6. Recurrence Relations. 

It would be tedious to give further analysis of these functions ; so, to conclude the 
paper, I shall merely state certain recurrence and other relations satisfied by these 

q 2n J q (m ,n-l, x 2 ) + J q (m ,n + l,x 2 ) = L^iAj^wi ,n,x 2 ), . (r?) 

x L 

q 2 '"J q (n , m - 1 , -c 2 ) + J,(» , m + 1 , x 2 ) = LJJaJ ,(m , n , x-) , 


whence we deduce 

<f" J g (m , n - 1) + J q (m , n + 1) _ [2w] _ (g 2n - 1) ($) 

q 2 '"J q (n , m - 1) + J q (n , m + 1) [2m] (q 2m - 1 ) ' ' 

q**J a (m ,n-l) + J q (m , n + 1) = ^ | J q '(m , n) + J^l3;\m , n) + x 2< ^lj q '"(m , n) + . ...'}, («) 
whence, in case q=l, 

J m ,»-i(.'-) + Jm,n+l(*) = — J'(™ ,»>«)' 

In the theory of Bessel's Function 

I a: eta J 

* Trans. R.S.E., vol. xli., pt, ii., p. 406'. 


The theorem analogous to this is 

U2„_l d y i_ 2)i CI \ I T m+n+r j _ m+n-T 

which in ^-functions is 

{(x 2, '->A)(z 1 --''A)} , x m+n+, J q (m,n + r) = x m+n J q (m,n). . . . (X) 

Other A equations are 

xL | s«+»J m . B J + arij- { a -+»+ 2 J,„ +1 , H+1 | = (2m + 2«) a f l +»J m> „ . . fc) 

| ce m+B Jj(m , n , x) 1 + ./-'A | a;'"+" +2 J g (m + l,n+l,a:)i= [2m + 2n]( - Y' + ' J,(m , n , gcc) . (v) 

A{a;" ,+ "J 4 (??i, w)}=a / "'+"- , </) 9 (?n ) «,a-), . . . . (p) 

<£,/?» , n , .<) + <£,(to 4 1 , n+ 1 , .'.) = ^~r n -^ g (m , n , qx) , 
from which, of course, various series may be obtained by repetition. 




( 283 ) 

XII. — The Evolution of the Eyebrow Region of the Forehead, with Special 
Reference to the Excessive Supraorbital Development in the Neanderthal 
Race. By Professor D. J. Cunningham, F.R.S. (With Three Plates.) 

(MS. received 23rd March 1908. Read 24th June 1907. Issued separately June 16, 1908.) 

One of the most striking features of the famous Neanderthal cranium consists in the 
strong projection which is exhibited in the glabellar and the supraorbital or eyebrow 
regions of the frontal bone. This character is rendered all the more important from 
the fact that all the specimens which have been collected since the Neanderthal cranium 
was discovered (1857), and which have been shown to belong to the same remote 
geological period, possess the same, or at least very much the same, remarkable 
prominence in the eyebrow region. These specimens are not very numerous, but, 
inasmuch as they represent the earliest remains of man with which we are acquainted, 
they possess a very special interest. In addition to the Neanderthal cranium, the group 
includes the two Spy crania, the Gibraltar skull, and the recently discovered Krapina 
remains. The Krapina remains are in a very fragmentary condition, but they 
apparently consist of portions of the skeletons of ten individuals, and the frontal bones 
all present the character in question. 

It is curious that, although all of the many observers who have written upon the 
Neanderthal race have dwelt upon the supraorbital projection and have recognised 
in it one of the leading peculiarities of the group, no one, with the exception of 
Schwalbe, has subjected the eyebrow region to a searching and critical examina- 
tion. In his recent important papers upon the so-called Pithecanthropus erectus 
and on prehistoric man, Schwalbe has thrown much light upon the value to be 
attached to the eyebrow projection, and has stimulated further research in the same 
field (5 to 11). * 

In the present investigation I have had the great advantage of having been 
afforded the privilege of studying the splendid collection of anthropoid and 
lower ape skulls in the British Museum. This privilege I owe to the kindness 
of Mr Oldfield Thomas, to whom I cordially offer my most grateful acknowledg- 
ments. To the British Museum specimens must be added the numerous anthropoid 
and lower ape crania in the Museum of the University of Edinburgh, which were 
also at my disposal. The full range of the investigation, in so far as the ape is 

* Professor Schwalbe has had the inestimable advantage of having recently been permitted to study at his leisure 
and in his own laboratory the Neanderthal remains. These he has described with great care (8), and has placed 
in tbe hands of other observers particulars in regard to these specimens of very great value. 




concerned, can best be appreciated by the following list of the specimens which 
have come under my notice : — 

Gorilla . 




10 young 
23 adult 

Orang . 


19 young 
27 adult 

Gibbon . 


(including 9 species) 

Semnopithecus . 


( „ 5 „ ) 

Nasalis . 




Cercopitliecus . 






(including 15 species 





Cebus . 


The large collection of human crania, including somewhere about 1500 specimens, in 
the Museum of the University of Edinburgh has more than sufficed for my purpose. 
I do not pretend to have examined all of these skulls from this point of view. I have 
chiefly directed my attention, for reasons that will be afterwards apparent, to the group 
of Australian crania (130 in number), and of these I selected the Victoria and Queensland 
specimens for special study. 

Degree of Projection of the Glabellar Part of the Frontal Bone. 

Schwalbe estimates the extent and degree of projection of the glabellar part of the 

eyebrow-region by measuring by the callipers the chords of the glabellar and cerebral 

curves or arcs of the frontal bone, and expressing the former as a percentage of the 

latter, thus : 

Gla bellar chord x 100 
Cerebral chord 

When dealt with in this way, the Neanderthal cranium gives an index of 44*2, and the 
Spy cranium No. 1 an index of 41*5. According to Schwalbe, the index in recent 
man rarely reaches 30, and varies between the limits of 21*4 and 31 "8 (8, p. 29). 

A New South Wales Australian cranium (xxix. B. l) in the ethnological collection 
of the University gives an index of 307, and another Australian skull from the Eiverina 
district (xxix. B. 12), with a still more prominent glabellar region, yields an index of 34. 
But I believe that even this index may be exceeded. Recently I received from Dr W. 
Ramsay Smith, of Adelaide, the head of an aboriginal Australian named Boco, in which 
there was an excessive development of the glabellar and supraorbital regions of the 
forehead (2). It had been carefully preserved by formalin injection, and measured 
over the soft parts the index reached the high figure of 52*3. Of course this cannot 


be taken as being comparable with those indices obtained from measurements of the 
skull, but there is reason to believe that if the calvaria of Boco were denuded of 
the scalp covering it would give an index at least as high as that of the Neanderthal 

But what is the value of this glabello-cerebral index of Schwa lbe ? Can we rely 
upon it giving a true and proper idea of the relative extent and degree of projection of 
the pars glabellaris of the cranium ? I do not think that we can, and I look upon the 
figures given above as being of little value, and in certain respects misleading. If the 
mesial length of the frontal bone, measured either by the tape or the callipers, from the 
nasion to the bregma were relatively (even approximately) the same in different skulls, 
some reliance might be placed on the index ; but when we find in three skulls so similar 
in the degree of glabellar projection as the Neanderthal, Spy 1, and Eiverina the total 
frontal length, as ascertained by the tape, so very different as 133 mm., 120 mm., and 
147 mm. respectively, it is evident that in these cases the index expresses variations in 
the length of the cerebral part of the frontal bone more than variations in the length of 
the glabellar part. Indeed, no index is necessary. The proper comparison to institute 
is one between the absolute measurements of the glabellar part made by the tape over 
its curvature from nasion to ophryon in different skulls. When this is done in the 
crania under consideration we obtain the following figures, and from these we can best 
realise the extent and degree of glabellar projection : — 

New South Wales cranium (xxix. 

B. 1) . 

. 31 mm. 

Spy No. 1 

• 40 „ 

Riverina skull (xxix. B. 12) 

■ 41 „ 

Neanderthal cranium 

■ 43 „ 

Boco (over scalp tissues) 

• 50 „ 

We are thus enabled to conclude that the high degree of glabellar development which 
is seen in the Neanderthal group, and which in it constitutes a distinct and definite racial 
character, may nevertheless be attained as an occasional variation in certaiu individuals 
of other races — notably the Australian race. 

GtIUFRIDa-Ruggeri (4), in a suggestive and well-reasoned paper, likewise expresses his 
dissatisfaction with Schwalbe's glabello-cerebral index. He refers to the variability in 
the position of the bregma due to differences in the form of the coronal suture,* and 
then he goes on to remark : " Even the position of the nasion varies according as the 
fronto-nasal suture is semilunar, triangular, or horseshoe-shaped, and this exercises an 
influence on the length of the glabellar chord. I have obtained higher indices in certain 
Melanesian skulls in which the bregma was not much displaced backwards." He gives 
five examples in which the index varied from 30*2 to 33*3, and then remarks : " Finally, 
in the skull 760, which is one of the most interesting in the Anthropological Museum in 
Rome, I obtained an index of 39*7, to which I would draw the attention of Professor 

* This is merely another way of stating that the variations in the position of the bregma are due to variations in 
the extent of the frontal field of the cranial wall. 


But while the glabellar region in individual cases in recent man may assume pro- 
portions quite as great as those which are seen within the Neanderthal group, I would 
not have it supposed that I attach undue importance to the agreement in this respect. 
Certain suggestive and significant points of difference will be brought out in the course 
of this paper. To one of these we may refer at the present time. The depression above 
the glabellar and supraorbital regions in the Neanderthal and Spy crania is much more 
extensive than in the case of any recent skull or in the skulls of any other prehistoric 
race. This constitutes a marked and highly important distinction, and one which brings 
the Neanderthal type of cranium into closer relation with that of the chimpanzee and 
the gorilla. Schwalbe fully recognises the significance of this character. The forward 
expansion of the cerebral part of the frontal bone in response to the increased develop- 

Fig. 1. — This figure is reproduced from Schwalbe's article upon 
"Das Schadelfragment von Briix, etc." (p. 109). It is an 
outline representation of the supraorbital region of a mandrill 
(Oynocephaliis mormon). 

a. Fossa supraglabellaris. c. Groove between b and 

b. A reus superciliaris. d. Arcus supraorbitals. 

ment of the brain in recent man has to a large extent obliterated this highly suggestive 
cranial feature in the Neanderthal forehead. 

Form and Morphological Constitution of the Eyebrow Region. 

Schwalbe has given a large amount of attention to the form and morphological con- 
stitution of the eyebrow region of the forehead in the Neanderthal race (6, 8, 10), and 
although I cannot accept certain of the conclusions at which he has arrived, it should be 
recognised that in this field of work he has broken new ground and has added greatly to 
our knowledge of the evolution of this part of the skull. He holds that the supraorbital 
region in the Neanderthal group is distinctive of that race, and that it differs in form 
and mode of construction from the same region in any other race, either past or present. 
He elaborates this point with much ability in several of his writings, and gives this 
feature a leading place amongst the characters which are peculiar to the crania of that 
primitive group. 

In briefly stating the views of Schwalbe, I shall follow the description which he gives 


in his recent paper on " Das Schadelfragment von Briix und verwandte Schadelformen " 
(10). He takes as a type the condition presented by the frontal bone of the mandrill 
(Cynocephalus mormon), and he recognises, above the nasion and the margin of the 
orbital opening, two regions, viz. (l) the superciliary ridge (arcus superciliaris), which 
extends upwards and outwards from the glabellar swelling (fig. 1, b); and (2) a lateral 
area situated below and to the outer side of the superciliary ridge. This area he terms 
the arcus supraorbi talis (fig. 1, d). It stretches from the supraorbital notch to the 

Fig. 2. — The frontal bone of an Alsatian, with strongly marked superciliary eminences (from 
Schwalbe, " Das Schadelfragment von Briix, etc.," p. 110). 
a. Fossa supraglabellaris. d. Trigonum supraorbital. 

6. Superciliary eminence. c. Groove between 6 and d. 

external angular process of the frontal bone, and forms a three-sided field between the 
superciliary ridge above and the margin of the orbital opening below. 

Schwalbe further states that the same parts, in very much the same relationship to 
each other, may be seen in the eyebrow region of recent man, and he gives outline 
sketches of the frontal bone of a mandrill and of an Alsatian to illustrate this point. 
These figures I have taken the liberty to reproduce, so as to make his position on this 
matter absolutely clear (figs. I and 2). In both of these drawings it can be seen that there 
is an arcus superciliaris (b) and an arcus supraorbitalis (d), and that these are separated 
from each other by an oblique furrow, the sulcus supraorbitalis (c), which ascends from 
the supraorbital notch in an outward and upward direction. 



The arcus supraorbitalis (d) by its lower border forms the greater part of the 
upper portion of the margin of the orbital opening, and in the recent human skull, 
according to Schwalbe, it presents a depressed or flattened field. For this reason 
he suggests that it should be termed the planum or trigonum supraorbitale. The latter 
name, in the vast majority of recent human skulls, admirably expresses its general 

To the inner side of the supraorbital notch the margin of the orbital opening is 
formed by the inner portion of the superciliary ridge (arcus superciliaris) as it runs into 
the glabella. The supraorbital notch, therefore, is an important landmark ; to its inner 
side the orbital margin is formed by the superciliary ridge ; to its outer side it is 
formed by the lower border of the trigonum supraorbitale. 

Fig. 3. — Outline sketch of the supraorbital region of the Neanderthal cranium, taken from 
Schwalbe's article on " Der Neanderthalschadel," p. 11. 

a. Slight notch (supraorbital notch (?) ). e. Eminentia supraglabellaris. 

b. Slight elevation on the left supra- /. Median frontal ridge. 

orbital border. g and h. Foramina supratrochlearia. 

c. Glabella, k. Notch on right supraorbital border 

d. Fossa supraglabellaris. (supraorbital notch). 

Such is Schwalbe's account of the supraorbital region in recent man and in all 
primitive races, with the single exception of the Neanderthal race. There cannot be a 
doubt that the description accurately conveys the condition which is present in the 
majority of recent skulls, and that it reproduces the type which exists not infrequently 
in the young mandrill and many other different forms of ape. But Schwalbe falls 
into error in asserting (l) that this is the only form of supraorbital region which 
exists in recent man ; and (2) that the form of the eyebrow region which is seen 
in the Neanderthal, Spy, and Krapina crania is never met with in the crania of 
the present day. 

In the Neanderthal cranium the supraorbital region is developed in the form of a 
strongly projecting continuous arch, which extends from the glabella to the external 
angular process (fig. 3). I am in complete agreement with Schwalbe as to the consti- 
tution of this arch. The superciliary eminence and the trigonum supraorbitale have 



become fused together so as to produce the striking arcuate elevation which distinguishes 
the eyebrow region in this cranium. Further, the term, torus supraorbitalis, which 
Schwalbe suggests should be applied to it appears to me to be both useful and 
appropriate. Indeed, I only take exception to the assertion that, while in the Neander- 
thal cranium the two elements, the arcus superciliaris and the trigonum supraorbital, 
run together and become fused into one continuous arch, these elements invariably 
remain separate in the crania of recent man. 

Fig. 4. — Outline tracing of the frontal region of a French skull. 
b. Arcus superciliaris. d. Trigonum supraorbitale. 

In studying the eyebrow region in man and the ape it will be convenient to look 
upon the elements which are typically present as being three in number. These 
elements are represented by the following parts : — 

1. The supraorbital margin or the frontal part of the rim of the orbital opening. 

2. The superciliary arch or ridge. 

3. The trigonum supraorbitale. 

According to the manner in which these three elements are arranged with reference to 
each other, three types of the supraorbital region may be distinguished. 

At the same time, however, it should be recognised that the three elements are not 
always present, either separate and distinct from each other, or in combination with each 



other. There are certain ape crania in which the arcus superciliaris is absent, and in every 
large collection of human crania a few specimens will be found in which the same deficiency- 
may be observed. In the latter the glabella and the region above the orbital opening 
are flat and vertical, and similar in appearance to what is seen in the forehead of the 
European child before the superciliary ridges and the frontal air sinuses are developed. 
This form of supraorbital region would seem to occur most frequently in certain 
African races. The only skull in the University collection in which there is absolutely 
no trace of a superciliary eminence or of a glabellar fulness is that of a male Nupe from 

Fig. 5. — Frontal region of a Kham warrior from Thibet (Museum, No. xxiv. A. 2). 
b. Superciliary eminence. d. Trigonum supraorbital. 

Nigeria which I received from my friend and former pupil, Dr Howard Ensor. There 
are several specimens (e.g. an Andaman skull, the skull of an adult Negress, a West 
African skull presented to me by my colleague Professor H. Littlejohn, the skull 
of a Ba-Mbala native given to me by Mr E. E. Torday, etc., etc.) which at first 
sight appear to be devoid of glabellar and superciliary eminences ; but when these 
specimens are inspected in a proper light it is noticed that there is a general 
fulness in this region, and perhaps a scarcely perceptible indication of the arcus 
superciliaris. In all there is an appreciable flattening in the region of the trigonum 


Type I. — There are many human crania in many different races in which the three 
elements of the supraorbital region are distinct and separate. The same also may be 
observed in certain species of ape. 

The supraorbital margin extends in a continuous and arch-like manner from the 
fronto-malar suture on the outer side to the fronto-maxillary suture on the inner side. 
It is divided into a long outer part and a shorter inner portion by the supraorbital 
notch. The outer portion is a projecting and sharply defined edge ; the inner part, 
quite distinct from the superciliary arch, sweeps downwards immediately in front of 
the trochlear pit, and although much less prominent than the outer part is yet quite 
obvious (fig. 4). 

The superciliary ridge or arcus superciliaris is a semilunar, smooth elevation which 
lies above the inner part of the inner portion of the supraorbital margin (fig. 4). Its 
inner end curves downwards into the glabella, and is separated from the corresponding 
eminence of the opposite side by a narrow, shallow median depression ; its outer end 
fades away as it approaches the trigonum supraorbitale. 

The trigonum supraorbitale lies to the outer side of the superciliary eminence. It 
is a triangular depressed field, with its apex at the fronto-malar suture, which is included 
between the outer part of the margo supraorbitalis and the anterior prominent part of 
the temporal ridge. 

Fig. 4 is taken from a lecture specimen of the frontal bone which I am in the 
habit of using for class purposes. Fig. 5 is the frontal bone of a Kham warrior from 
Thibet, in which Type I. of the supraorbital region is particularly well marked. Owing 
to the somewhat unusual length and strongly curved character of the superciliary 
arches, the supraorbital region in this skull presents a striking appearance. The 
trigonum supraorbitale is very depressed. 

This type of eyebrow region does not appear to be distinctive of, nor indeed more 
frequently present in, any one race more than another. It occurs in most, if not in 
all, recent races, and even amongst Australian skulls it is not infrequently met with, 
as will be seen from the following figures : — 

Number of Times present in a Group of 25 Australian Skulls. 

Natives of Victoria — 8 males . . . 1 • 

5 females . . 2 

Natives of Queensland — 10 males ... 1 

2 females . . 

In this small group of Australian skulls, therefore, Type I. of the eyebrow region 

occurred four times, or in 16 per cent, of the specimens. 

This type of the supraorbital region is also seen in certain apes, although probably 

in no case is it to be regarded as distinctive of any particular species. Still, it should 

be noted that it was present in a very definite manner in each of the three specimens of 

Macacus cy clops which I had the opportunity of studying. 



The specimens in which it occurred were the following :— 

Macacus cyclops . . . .3 times 

Macacus nemestrinuts 
Macacus rhesus 
Macacus speciosus 
Cynocephalus niger 

Cynocephalus porcarius . . 1 ,, 

The skull of a young specimen of Macacus cyclops is figured in PL I., fig. 14. It 
exhibits a somewhat unusual condition, and the supraorbital region in two other skulls 
belonging to the same species presented a similar appearance. A slightly raised oval 
field above and behind the glabella and inner part of the supraorbital margin represents 
the superciliary eminence. It is quite isolated, and stands apart from the glabella 
and the margo supraorbitalis. 

The skull of an older specimen of macaque monkey (Maxacus rhesus) is seen in 
PL I., fig. 15. In this specimen the three elements of the eyebrow region are also distinct 
from each other, but the superciliary projection is very different in form and in its 
degree of prominence. It forms a pronounced, elongated eminence which runs out- 
wards above the margin of the orbital opening, and presents a different texture from 
that of the surrounding bone. Its outer extremity runs into the trigonum supra- 
orbitale and reaches the anterior part of the temporal ridge, with which it in part fuses. 
Throughout its whole length it is separated from the margo supraorbitalis by a strongly 
marked groove, whilst its inner end remains distinct and does not run into the 
glabella nor fuse with its fellow of the opposite side. 

A somewhat similar arrangement is seen in the skull of an adult specimen of the 
black ape of the Celebes (Cynocephalus niger). The superciliary eminence is in the form 
of a sharply marked ridge placed above the margin of the orbital opening, and separated 
from it by a deep sulcus. Its outer end remains free and ends in the trigonum supra- 
orbitals In the median line there is a small rounded prominence which lies between 
the inner ends of the two superciliary ridges, and probably represents the glabella. 

Tn the young chacma baboon (Cynocephalus porcarius) we sometimes meet with a 
condition which presents a strong resemblance to what is seen in cases where Type I. 
of this region is present in the human skull (PL I., fig. 17). The strongly marked 
superciliary eminences are semilunar in outline and curve upwards and outwards from 
the region of the glabella to the temporal ridge on either side. The eminences remain 
distinct from each other, but turn downwards into the glabella. A broad, shallow groove 
separates the superciliary ridge from both the inner and outer parts of the margo 

It must be clearly understood that I do not put forward these specimens, all of 
which fall clearly within the limits of Type I., as being characteristic of the species of 
ape to which they belong. As in the case of the same form of eyebrow region in the 


human skull, they must be looked upon as being merely peculiar to certain individuals 
and as occurring sporadically in several different genera and species of ape. 

Anyone who studies the ape skull can readily satisfy himself that within one species 
several different forms of the supraorbital region may be encountered. To some extent 
this is apparently the result of the vagaries of individual development, but it is probably 
more frequently due to changes which occur with the advance of age. As adult life is 
approached, there appears to be a tendency towards a partial or complete fusion of the 
three elements, and thus amongst the apes the same individual may present very 
different types of eyebrow region at different periods of life. 

Type II. — In Type II. are included those skulls which exhibit that condition of the 
supraorbital region which Schwalbe has described as being peculiar to recent man, and 
also to the mandrill and other species of ape. The superciliary projection has coalesced 
with the part of the supraorbital margin which lies to the inner side of the supraorbital 
notch. On the outer side of the notch the eminence extends outwards with a varying 
degree of prominence and for a varying distance towards the trigonum supraorbitale. 
From the latter it is separated by a faint groove which ascends obliquely upwards and 
outwards from the supraorbital notch. The trigonum supraorbitale, which varies in its 
extent according to the degree of development of the arcus superciliaris, is, as a rule, 
depressed and flattened. This form of the supraorbital region occurs in all races, past 
and present, with the exception of the Neanderthal race, and is undoubtedly the con- 
dition which is most distinctive of man. 

In the group of 25 Australian skulls to which I have already referred, Type II. of 
the supraorbital region occurred no less than 18 times, or in 72 per cent. 

Natives of Victoria — 8 males . . .5 times 

5 females . . . 3 ,, 

Natives of Queensland — 10 males . . . 8 „ 

2 females . . . 2 ,, 

As Schwalbe has shown, Type [I. of the supraorbital region also appears in the 
mandrill (6 and 10), but the form which he has figured (fig. 1, p. 286) only occurs 
in young specimens. In the adult skull there is a tendency towards the coalescence of 
the different elements and the formation of a torus which is morphologically equivalent 
to what is seen in the gorilla and the chimpanzee. Still, this is not by any means the 
invariable result of advancing age in the mandrill. In the largest and most character- 
istic skull of this ape in the British Museum, the form presented by the supraorbital 
region falls clearly within Type II., although the massive superciliary eminence is 
restricted to the inner part of the region and does not stretch outwards in the form 
of an elongated ridge, as in the young specimens of the same species. 

It is in the genus Cynocephalus that we find the closest approximation to Type II. 
as it is exhibited in the human skull. Probably this is the only group of apes in which 
this type of eyebrow region is almost invariably present. Young specimens may be 
found in which Type I. occurs (PL I., fig. 17) ; these are rare. Again, as age advances 



there is undoubtedly a tendency for the elements of the region to fuse together to form 
a torus ; but in the fifty specimens I have examined I have not seen a skull in which 
the coalescence is complete. There is always a trace of the groove which intervenes 
between the superciliary eminence and the trigonum supraorbitale. Figs. 18 and 19, 
PI. I., represent the usual appearance which is presented in this genus. In the middle 
line the two superciliary arches are completely fused to form the glabellar eminence, 
and here also they coalesce with the part of the orbital margin which lies to the inner 
side of the supraorbital notch. From the glabella the two superciliary ridges arch 
outwards like the two limbs of the letter Y, and, gradually tapering, each ends in the 
neighbourhood of the temporal ridge. A strongly marked groove separates the super- 
ciliary eminence from the trigonum supraorbitale. The term "trigonum" is hardly 
appropriate for the part of the region which lies below and to the outer side of this 
groove. It is true that it is a somewhat triangular area, but it is not flattened and 
depressed as is usually the case in the human skull ; it is prominent, ridge-like, and 
highly curved in accordance with the curvature of this part of the orbital opening. 
This may be regarded as an approach to the condition termed by Schwalbe a " torus 
supraorbital. " By the disappearance of the intervening groove and the consequent 
coalescence of the superciliary eminence and the trigonum supraorbitale, the form 
distinctive of the Neanderthal race, the gorilla, chimpanzee, etc., would be produced. 

The condition seen in the New Hebrides skull figured in PI. III., fig. 23, presents a 
striking resemblance to the form of supraorbital region which we have described as 
distinctive of the baboon. Two points of difference, however, are noticeable. The two 
superciliary arches have drawn away from each other in the glabellar region, and their 
inner ends are separated from each other by a shallow median groove. This is due to 
the broadening of the forehead in man, and the coincident widening of the glabellar 
region and of the interval which separates the orbits. In the baboon the narrow fore- 
head is associated with a narrow glabella and a narrow root of the nose, and thus the 
superciliary ridges are brought together in the middle line. The second point of 
difference is in the trigonum supraorbitale. This area is flat and depressed in the New 
Hebridean skull, although there is seen a tendency for the outer part of the supra- 
orbital margin to form an arched prominence as in the baboon. 

Although not distinctive of any other group of apes, Type II. frequently occurs 
in individuals of other genera and species. I have noted its presence in the follow- 
ing specimens : — 

Semnopithecus (in two species, viz. aut 

Nasalis lanatus 

Cercopithecus . 


Mucacus (in eight different species) 

Presbytes maurus 

Calms (in four different species) 

In certain of the nine specimens entered in the above list as belonging to the genera 

ratus and cristatus) 

4 times 




. 5 


. 5 


. 14 




iforon in flifl ahn^TD lief qq 


"iolfinmn(r fr 

^ flip rrPTiPra. 


Semnopithecus and Cercocebus the fusion of the three elements of the supraorbital region 
was very nearly complete, but in the region of the trigonum the superciliary element 
showed as a distinct ridge above the outer part of the orbital margin. In these cases 
the supraorbital notch fails as a guide to the line of separation between the superciliary 
and other parts of the region. 

Schwalbe gives a careful and accurate account of the supraorbital region in the 
cranium of Pithecanthropus erectus, in so far as this can be studied in the plaster cast 
of the specimen, and he supplements his remarks by a drawing. I have had the 
advantage of being able to examine two casts of this cranium, one of which was 
presented to me by Dr Eugene Dubois when he visited Dublin, and another which he 
gave to Sir William Turner. Both of these specimens are in the Anatomical Museum 
of the University of Edinburgh. They are sharp and excellent casts, and they agree 
in every particular with Schwalbe's description. I am thus able to confirm Schwalbe's 
view that the eyebrow region presents the form which I have included under Type II. 
It should be noted, however, that the fusion between the superciliary and the supraorbital 
elements is very nearly complete, and the condition is one which closely corresponds to 
that which is frequently seen in the aged Anubis baboon. Indeed, from the appearance of 
the eyebrow region I think we may conclude with some degree of certainty that the 
cranium of Pithecanthropus erectus is that of an aged individual. 

As Schwalbe points out, it is an important and significant fact that it is to the 
baboon, much more than to the gorilla or chimpanzee, that Pithecanthropus erectus 
shows a resemblance in this respect (6). 

Type III. — In this form of the supraorbital region all the three elements which 
enter into its formation become fused together so as to constitute a continuous arch, 
the torus supraorbitalis, which bounds the orbital opening above and forms a varying 
amount of the fore-part of the roof of the orbital cavity. This arch extends from the 
glabella to the fronto-malar suture, and in its typical condition it shows no trace of the 
constituent elements of which it is composed, so completely have these become blended 
the one with the other. Of such a nature is the supraorbital region in the adult 
chimpanzee, in the gorilla, and in a large number, if not the majority, of the lower apes 
of the Old World. Still, it is not correct to suppose that even in the gorilla and the 
chimpanzee this is the only condition of the eyebrow region which may be present, and 
that there never occurs an approach to the form distinctive of Type II. In young 
specimens of the chimpanzee the superciliary eminence, although fused with the 
supraorbital torus and not marked off on its outer side by any distinct groove, is yet as 
a rule easily discernible, not only by the greater degree of prominence of this part of 
the arch, but also by a difference in its texture. The same also may be occasionally 
seen in the adult chimpanzee. In such cases the supraorbital notch, which Schwalbe 
takes as the demarcation between the supraorbital and superciliary elements, fails to be 
a true guide. The superciliary element extends outwards in the torus far beyond this 
point. In the beautiful illustrations of anthropoid crania which are given by Selenka, 


there is the figure of the skull of a young female gorilla in which the superciliary 
element of the torus supraorbitalis is distinctly seen (12, fig. 112, p. 102). It is still 
more evident in the skull of an adult male chimpanzee in the Anatomical Museum of 
the University of Edinburgh. The photograph of this skull is given in PI. I., fig. 15. 
Transition forms between Type II. and Type III. of the supraorbital region are thus 
met with in both the chimpanzee and the gorilla. 

Of the nineteen skulls of the adult chimpanzee which I had the opportunity of 
studying in the British Museum, three presented this transition form. In one (Museum 
No. 2, c. 1130) the condition was particularly strongly marked. The torus supra- 
orbitalis was divided into two distinct parts, the inner two-thirds, which contained the 
superciliary element, being very prominent as compared with the outer third.* This 
is to be looked upon as being merely the retention of a juvenile character. It is the 
rule in the young chimpanzee for the brow-ridges to begin to assume form at the time 
when the milk-teeth erupt. In the early stages it is generally possible to distinguish 
the superciliary element in the torus supraorbitalis ; but, as a rule, after the second molar 
tooth has made its appearance the superciliary eminence becomes completely blended 
with the other elements which build up the torus, and the supraorbital projection assumes 
its typical form. 

In the adult gorilla, the amalgamation of the different elements in the torus supra- 
orbitalis is so complete that little indication of the composite character of the latter can 
be detected. Still, when the arch is viewed from above a general fulness in the super- 
ciliary region points to the presence of the superciliary element. I have not had the 
opportunity of studying the condition in the skull of the young gorilla. 

The supraorbital region in the Neanderthal, Spy, and Krapina remains presents 
features which place it within Type III. As Schwalbe has shown, the different elements 
have become blended in a strongly projecting torus supraorbitalis, in which there is little 
or no indication of its composite character. This undoubtedly forms one of the leading 
features of the race, but Schwalbe is in error in concluding that the possession of an 
eyebrow region of this formation is the exclusive property of the Neanderthal group. 
Individuals are met with in recent races in whom the same formation may be seen in 
different degrees of development, and many specimens occur which present transition 
forms between Type II. and Type III. 

In all large collections of Maori and New Guinea crania individual specimens which 
belong to Type III may be found ; but in these cases the torus supraorbitalis, while show- 
ing a complete fusion of the superciliary and supraorbital elements and the formation of 
a continuous and uniform arch above the orbital opening, does not attain a great degree 
of projection, nor does it take much share in the formation of the orbital roof. 

Amongst the Australian crania the case is different. Occasional specimens will be 

* In another specimen (Museum No. 2a) the appearance presented by the supraorbital region strongly resembles 
the form usually seen in the aged baboon. A very evident groove (most evident on the left side) limits the super- 
ciliary element on its outer side. 


found which present a torus supraorbital very similar to that present in the Neanderthal 
and Spy crania. The skull in the University collection which shows this character in 
the most pronounced form is figured in PI. II., fig. 20. It is from a native of New 
South Wales (xxix. B. 1), and in general form and external configuration the eyebrow 
region exhibits a marked resemblance to the corresponding region in the Neanderthal 
cranium. A massive projecting bony arch, composed of the fused superciliary and supra- 
orbital elements, extends from the glabella to the fronto-malar suture. 

Upon the right supraorbital torus of the Neanderthal cranium there is an oblique 
furrow which extends upwards and outwards for a distance of about 12 mm. (fig. 3, s). 
This groove is visible on the cast ; and by both Schaffhausen and Virchow it was 
thought to be of traumatic origin. Schwalbe, however, has shown that, although not 
continuous with, it yet lies in the line of a notch on the supraorbital margin, around 
which the supraorbital or outer branch of the frontal nerve had no doubt turned in its 
passage to the forehead (fig. 3, k). This naturally suggests the possibility that the groove 
in question is not the result of a wound, but is the pathway of the nerve. Against this 
conclusion must be reckoned the position and direction of the furrow. It courses over 
the outer third of the torus supraorbital , and its direction is so oblique that, if continued 
outwards, it would strike the temporal ridge about 14 mm. above the fronto-malar suture. 
It would seem unlikely, therefore, that the supraorbital nerve or one of its branches would 
pursue such an initial course on reaching the forehead. Still, it should be noted that in 
the Neanderthal cranium the supraorbital notch is situated much further out than usual. 
It interrupts the margin of the supraorbital arch much nearer its outer than its inner end 
(viz. 27 mm. from the fronto-malar suture, and 35 mm. from the fronto-maxillary suture). 
That this represents the notch in question there cannot be a doubt, because Schwalbe 
has shown that on the roof of the orbit a distinct nerve-groove leads to it (fig. 6). 

I have alluded to these features in the Neanderthal cranium because we find on the 
right side in the cranium of the New South Wales native (xxix. B. 1) a condition which 
in some respects is similar. The supraorbital notch cuts into the supraorbital margin 
about its mid-point, and from this two sharply defined grooves pursue a short divergent 
course on the under aspect of the torus supraorbitalis (figs. 20 and 21, PL II.). 
Immediately above the outer of these furrows and in a line with it is a shallow oblique 
sulcus on the front face of the torus. This recalls the groove on the right torus of the 
Neanderthal specimen ; but its position is different. It does not lie so far out, and it is 
not so oblique in its direction. 

I am inclined to think that in both the Neanderthal and New South AVales crania 
the groove in question is of the same nature, although I am satisfied that in neither case 
was it the pathway of a nerve. I believe that it represents in both cases the line of 
fusion between the superciliary and supraorbital elements of the torus. In the Australian 
skull, where the groove is faintly marked, and also in the Neanderthal cranium, where it 
is strongly marked, the condition resembles the appearance which is presented by the 
torus in those chimpanzee crania in which the superciliary element is not completely 



blended with the supraorbital element. As has been already noted, a distinct groove 
is present in one of the British Museum specimens in this situation (Museum No. 2a). 

On each side and at precisely the same place the supratrochlear nerve grooves the under 
aspect of the inner part of the torus supraorbitalis of the New South Wales cranium 
(PI. II., figs. 20 and 21). In the Neanderthal cranium these nerves have been conducted 
to the forehead through two short canals which occupy precisely similar positions 
(fig. 6). It is interesting to note that, while the supratrochlear notches (when they 
exist) in the human skull show a great constancy in point of position, the supraorbital 
notches vary considerably in this respect. 

In all questions relating to the district of the frontal bone under consideration the area 
immediately above the glabella and the eyebrow eminences is of high morphological import- 

Fic. 6. — Orbital and nasal aspect of the Neanderthal cranium, from ("Der Neanderthal 

Schadel," p. 38, fig. 10, 1). 
e. Supraorbital notch. 

a. Supratrochlear foramen. 

c. Groove on roof of orbit for supratrochlear nerve. 

d. Groove on l oof of orbit for supraorbital nerve. 

b. Groove on roof of orbit for frontal nerve. 
/. Supratrochlear foramen. 

g. Groove for supratrochlear nerve. 
h. Groove for supraorbital nerve (?). 
1. Lachrymal fossa. 

ance. In this area Schwalbe recognises three parts, viz. a median portion which he terms 
the facies supraglabellaris, and two lateral depressed portions which he calls the sulci supra - 
ciliares. The width and extent of the sulci supraciliares in the Neanderthal cranium 
constitute two of the most pithecoid characters in the Neanderthal and Spy crania. The 
massive eyebrow projection is separated from the curved frontal plate which covers the 
frontal extremity of the cerebrum by a wide depression which in some degree resembles 
the extensive depressed area which occupies a similar position in the adult chimpanzee 
and gorilla. This condition is associated with a low and flattened forehead, and bespeaks 
a feeble development of the frontal lobes of the cerebral hemispheres. As these lobes 
assume the massive rounded form distinctive of recent man, the cerebral part of the 
frontal bone assumes a higher degree of curvature for their proper accommodation, and 
its lower part advances towards the eyebrow region, and thus tends to diminish the 


extent of the sulci supraciliares and also at the same time to tilt them forward, so that 
in the higher types of forehead their surfaces no longer look upwards as in the gorilla, 
but almost directly to the front. In this respect there cannot be a doubt that the 
Neanderthal cranium occupies an intermediate position between, say, the chimpanzee 
and the vast majority of recent skulls ; but occasionally a recent skull is met with 
which shows an approximation to the Neanderthal condition, and which, so to speak, 
occupies the gap which exists in this respect between the Neanderthal race and man of 
the present day. 

The New South Wales cranium is a specimen of this kind. It possesses tori supra- 
orbitales similar in form and constitution to those of the Neanderthal cranium, but it 
differs from the latter in the extent of the sulci supraciliares. These, it is true, have 
an unusual width and extent for recent man, being 10 mm. wide on the right side and 
13 mm. wide on the left side ; but they fall far short of the corresponding sulci in the 
Neanderthal cranium, which are 18 mm. wide on the right side and 19 mm. wide on the left 
side. Further, this broad sulcus in the Neanderthal cranium is more depressed, and its 
surface looks more directly upwards, than in the New South "Wales specimen. 

Associated with this approximation of the eyebrow region to the Neanderthal type, 
the New South Wales skull presents a low and degraded forehead. The bregma-nasion- 
inial angle is only 53°, and the index of the frontal curve is so low as 18 (3). 

Amongst Australian crania it is possible to find specimens in which the eyebrow 
region conforms in every respect with the requirements of Type III., and yet possess, 
associated with this, a degree of frontal curvature as bold as that which is characteristic 
of the European skull. In PI. II., fig. 22, there is reproduced the photograph of the skull 
of a male Queenslander. In this specimen the B.N.I, angle is no less than 64° — the 
average angle in the Australian being 60°, and in the Scottish 61°. Further, the index 
of the frontal curvature in the Queensland skull in cpiestion reaches the high figure of 
24*5 (3). In this skull the eyebrow ridges form two continuous arches as in the 
Neanderthal cranium, but the supraciliary depressions are narrow grooves and look almost 
directly forward. 

The two Australian skulls which I have selected as examples of Type III. of the 
supraorbital region exhibit this condition in a more marked degree than any other 
specimens in the ethnological collection of the University. In the twenty-five 
Australian skulls specially examined, Type III. occurred in three cases, all of which 
were males, viz. in two natives of Victoria and in one native of Queensland. I have 
not seen this type of supraorbital region in any female skull. 

Amongst the Australians, transitional forms of the eyebrow region between Type II. 
and Type III. are very common. One of the most remarkable instances of this is to 
be seen in the skull from the Riverina district to which reference has already been 
made (xxix. B. 12). In this specimen the eyebrow eminences are very pronounced 
(PI. III., fig. 24). As mentioned before, the glabellar part of the frontal bone 
measures over the curvature 41 mm., or only 2 mm. less than the correspond- 



ing part of the Neanderthal cranium. The eyebrow projections form an almost 
continuous arch from the glabella to the extremity of the external angular process of 
the frontal bone ; still, the line of fusion between the superciliary and supraorbital 
elements is clearly indicated by a faint groove. There is no trigonum supraorbitale. 
A specimen such as this is very instructive, because it shows that the arched eyebrow 
elevation (the torus supraorbitalis) distinctive of Type III. is not merely formed by a 
strengthening of the superciliary eminence and by its fusion with the supraorbital part, 
but also by a great development of the latter as well. These changes lead to an oblitera- 
tion of the trigonum supraorbitale — the feature which Schwalbe considers distinctive of 
all recent skulls. This increased development of the supraorbital element is also evident 
in the baboon, although the projecting arch-like elevation which it forms is sharply cut 
off from the superciliary eminence by a pronounced sulcus. The study of the ape skull 
as well as a large series of human crania thus renders evident the steps by which the three 
different types of the supraorbital region have been evolved, and more especially do 
we see the manner in which Type II., which is chiefly characteristic of recent man, 
may be transformed into Type III., which is distinctive of the gorilla, the chimpanzee, 
the Neanderthal race, and also of a few exceptional individuals of the present day. 

Supraorbital Region in the Orang and the Gibbon. 

Had our survey been extended over the entire range of the order Primates, it would 
have been necessary to have recognised other types besides the three which we have 
stated include all human skulls and the majority of ape skulls. The condition in the 
orang, for example, is interesting in so far that there is no evidence of the presence of 
a superciliary element. The supraorbital elements gradually develop into a projecting 
rim for the upper part of the orbital opening. In the young skull this makes its 
appearance about the same time as the milk-teeth. At first it is very feebly expressed 
and separated from the cerebral part of the frontal bone by a narrow groove. Towards 
the glabella it is extremely narrow, but as it is traced outwards it broadens, and as age 
advances it becomes more and more projecting and rough, until ultimately in the adult 
its outer part occupies the whole field of the trigonum supraorbitale. It presents the 
appearance at this stage as if it were something superadded to the frontal bone, and 
might be compared to the upper segment of an oval frame attached to the frontal 
portion of the margin of the orbital opening. Inasmuch as it does not contain a super- 
ciliary element, it is not a true torus supraorbitalis, and further, it does not take a large 
share in the formation of the roof of the orbit. 

In certain aged specimens of the orang skull the narrow inner end of this supra- 
orbital projection may be seen to expand in the glabellar region into a rounded knob- 
like eminence. Such cases are not common, and they suggest the possibility of this 
expansion being the representative of a superciliary eminence. I do not think, how- 
ever, that it can be reckoned as such. 

There are some forms of gibbon in which it is impossible to detect a superciliary 



element, and in which the supraorbital region presents a close resemblance to the 
condition present in the orang. In most cases, however, there is a strong tendency 
exhibited for the inner end of the supraorbital arch to swell out into a bulbous ex- 
tremity at the glabella. This in all probability is the superciliary element ; indeed, it 
may sometimes be seen to partially disengage itself from the supraorbital arch and 
present a condition which may be classed under Type II. 

Relation of the Supraorbital Region to the Orbit. 
Schwalbe rightly lays stress upon the fact that in the majority of apes the supra- 

Pre-cerebral part of the 
roof ol the orbit. 

Fig. 7. — Tracing from a sagittal section through a frozen male head (Irish), in a 
plane corresponding to the mid-point of the supraorbital arch (reduced). 

orbital eminence, which juts forward in a shelf-like manner from the front of the true 
brain-capsule, forms a considerable part of the roof of the orbit. The degree of promi- 
nence to which this condition may attain varies in different groups, and may be 
demonstrated by making a sagittal section through the frozen head in a plane which 
passes through the mid-point of the supraorbital arch. This method has the further 
advantage of showing the relation which this pre-cerebral part of the roof of the 
orbit presents to the eyeball. 

In the adult male gorilla the torus supraorbitalis may form as much as the anterior 
half of the orbital roof, and a similar relationship may also be seen in a pronounced 
form in the chimpanzee, baboon, and other forms of ape. This character is strictly 
correlated with the extent of the area to which we have applied Schwalbe's term of 



sulcus supraciliaris. As the brain advances in its phylogenetic growth, it reduces the 
width of this sulcus, encroaches upon the eyebrow projection, and extends the area of 
the orbital plate of the frontal bone which intervenes between the cerebrum and the 
orbital cavity. 

At the same time it should be noted that even in the European a considerable part 
of the orbital roof may be formed by the supraorbital projection of the frontal bone. Id 
fig. 7 is shown a sagittal section through the head of an Irish subject in which the eye- 
brow region was somewhat prominent. It will be seen that very nearly one-third of the 
orbital roof is formed by the under surface of this part of the frontal bone. Further, it 
will be noted that three-fourths of the eyeball lie in front of the brain and under shelter 
of this portion of the bone. 

cerebral part of the 
roof of the orbit. 
,- Arcuate eminence 
on face. 

Pre-cerebral part of 
roof of the orbit. 

Fig. 8.— Sagittal section through the frozen head of a 
young male chimpanzee, in a plane to the outer 
side of the mid-point of the supraorbital arch. 

Ficj. 9. — Sagittal section through the frozen head of an 
adult gibbon {Eylohatcs agilis) in a plane corre- 
sponding to the mid-point of the supraorbital 
arch. Note the large share which the torus supra - 
orbitalis takes in the formation of the roof of the 

Those familiar with the facial aspect of the gorilla and chimpanzee know that the 
torus supraorbitalis is not only concerned in forming a large part of the wall of the 
orbit, but also in producing a strong arch-like projection which juts forward on the face 
above and in front of the eyeball. Fig. 8 was obtained from a tracing of a sagittal 
section of a young chimpanzee (probably about three or four years old) ; the torus 
is thus far from having attained its full degree of development. Still, the manner in 
which it forms the facial feature referred to is manifest. In one respect this 
section cannot be compared with the other sections shown in figs. 7 and 9. The 
impression which it gives of the relation presented by the roof of the orbit and the 
brain to the eyeball is somewhat misleading. It will be noticed that the section 
has been made through a different plane. Had it passed through the same plane 
as in the case of the human head, the relation presented by the eyeball to the roof 
of the orbit and the brain would not have been markedly different from what we 



see in the other figures (see fig. 1, PI. X., Cunningham Memoir No. 2, Royal Irish 
Academy, 1886). 

Instructive tracings of the frontal bone, which show in some measure the relation of 
the torus supraorbitalis to the roof of the orbit, are given by Schwalbe for the Neander- 
thal cranium and by Sollas for the Gibraltar cranium. In these the outline of the 
posterior surface of the bone is not represented, and thus it is not possible to estimate 
the precise extent of orbital wall which is formed by the torus. Still, by these tracings, 
and also by an examination of the cast of the Neanderthal cranium, it is evident that in 
this respect the Neanderthal race presents a very marked approach to the pithecoid type 
On this point I am in complete agreement with Schwalbe. In no recent human cranium 
is the orbital roof so largely formed by the supraorbital projection of the frontal bone. 

Fig. 10. — Sagittal section through the mid-point of 
the supraorbital arch of the New South Wales 
cranium (xxix. B. 1). 

Fig. 11. — Tracing of the frontal bone of the Neanderthal 
cranium at the mid-point of the supraorbital 
arch. (From Schwalbe, "Zur Frage der Abstain- 
mung des Menschen," fig. 4, p. 22.) 

Nevertheless, specimens are met with which present some similarity to the Neander- 
thal cranium in this character. The New South Wales cranium (xxix. B. l) to which 
we have so frequently referred is a case in point. A section through the frontal bone of 
this specimen in a sagittal plane corresponding to the mid-point of the torus supra- 
orbitalis is shown in fig. 10, and when this is compared with Schwalbe's tracing of 
the same region in the Neanderthal cranium (fig. 11) a decided resemblance is apparent. 
To Schwalbe's tracing I have added by a dotted line what may be regarded as indicating 
the approximate position of the posterior surface of the frontal bone, or in other words 
the outline of the cerebrum. If this has been accurately represented (and I think that, 
when it is contrasted with the drawing of the New South Wales specimen, it will be 
admitted that it cannot be very far wrong), the maximum length of the pre-cerebral 
part of the roof of the orbit in the Neanderthal is 20 mm., and in the New South Wales 
cranium 16 mm. 


Relation of the Superciliary Eminence to the Frontal Air-Sinus. 

That the degree of prominence of the glabellar and supraorbital regions of the human 
skull is not necessarily determined by the degree of development of the frontal air-sinus 
is proved by the writings of Bianchi (l), Zuckerkandl (15), Logan Turner (14), 
and Schwalbe (6, 8, 10). Further, the topographical independence of the area repre- 
sented by the air-sinus and that occupied by the superciliary eminence is a well- 
established fact. Two questions, however, of much difficulty remain to be considered : 
( 1 ) Can any morphological connection be established between the superciliary eminence 
and the frontal air-sinus ? and (2) if no such relationship exists between these two 
factors, what is the morphological significance of the superciliary eminence, and how can 
its presence in so large a number of the members of the primate group be explained ? 

In considering these problems it should first be noted that where the superciliary 
eminence fails the air-sinus is also usually absent. We may take the orang as an 
example of this. As we have noted, there is apparently no morphological equivalent of 
the superciliary eminence in the orang. It is not correct to say, however, that in this 
animal the frontal air-sinus is never developed. In the British Museum I had the 
opportunity of making a tracing of the bisected skull of an adult orang in which a small 
sinus was present. 

But the absence of the sinus in cases in which there is no superciliary eminence 
proves very little. On the other hand, we are confronted with the fact that there are 
numerous catarrhine apes in which the superciliary eminence is highly developed, and 
yet in which there is not a trace of the sinus. Further, it should be borne in mind that, 
as Dr Logan Turner has so clearly demonstrated, the sinuses are not infrequently 
absent in individuals of apparently all races of man. According to this authority, they 
are absent in 7*5 per cent, of European skulls. 

It would almost appear, therefore, that there is no morphological connection between 
the superciliary eminence and the frontal air-sinus, and yet when we make a vertical 
section through the region in one of the lower apes (as, for example, the baboon or the 
macaque) we see that the eminence is due to a separation of the two tables of the frontal 
bone and the replacement, between them, of the ordinary diploe by open cancellous tissue. 
Indeed, the condition is identical to the changes which occur in the young human frontal 
bone preparatory to the extension into the region of the nasal cavity to form the frontal 
air-sinus (fig. 1 2). No doubt this is suggestive, but I am afraid we cannot conclude from 
this fact alone that any clear connection exists between the condition present in these 
apes and the subsequent step which leads to the formation of an open air-sinus in certain 
of the anthropoids and man. Still, it is just possible that the condition may indicate in 
these apes the phylogenetic step by which the sinus formation has been reached. 

If, then, as seems likely, there is no morphological connection to be traced between 
the superciliary eminence and the frontal air-sinus, how can we account for the presence of 
the former ? It is not required for the formation of an efficient torus supraorbitals ; 


this can be formed without its aid. As we have seen, the superciliary eminence 
frequently appears as an isolated eminence, standing absolutely independent of, and 
apart from, the other elements of the region. I suspect that this problem must remain, 
for the meantime at least, unsolved. It may be that, by adding to the volume of the 
supraorbital projection, the superciliary eminence increased the aspect of ferocity, which 
is generally associated with projecting brows, and thus contributed to the face a feature 
which would no doubt have been a decided advantage in those early struggling days. 
Little gain, however, is to be acquired by following a line of thought so highly 

As is well known, the torus supraorbitals in the gorilla and the chimpanzee is 
chiefly composed of dense bone. The frontal air-sinuses are relatively small, and are 

... Superciliary eminence 
just beginning to 

Frontal air-sinus extend- 
ing upwards from the 
nasal chamber into the 
superciliary region. 

Fig. 12. — Sngittal section through the frontal bone of a child, in the 
region of the superciliary eminence. 

situated at the base of the torus, close up against the inner table of the cranial wall. 
Further, they are chiefly confined to the glabellar region, and do not extend for any 
distance outwards into the superciliary part of the torus. Much interest is therefore 
attached to the observation of Professor Schwalbe (8) that in the Neanderthal cranium 
the air-sinuses lie well back, and that a thick layer of condensed bone forms their anterior 
wall. It is usual to find a similar pithecoid condition in the supraorbital region of 
the native Australian. The frontal air-sinuses, in this race, are as a rule relatively 
small, and they lie behind a mass of condensed bone. Thus Dr Logan Turner (14) was 
only able to map out these sinuses by the illumination method in less than a third of 
the Australian skulls he examined (in 20 out of 69 skulls), and in a very large number 
(30*4 per cent.) he found both sinuses totally absent. Amongst the Maori skulls, which, 
as we have noted, also show a tendency to Type III. of the eyebrow region, the same 
observer states that both sinuses were absent in 37 per cent, of the specimens he examined. 
These are striking facts, and undoubtedly point to an important affinity in this 


respect between the native Australians and the Neanderthal race on the one hand, 
and of both of these to the chimpanzee and the gorilla on the other hand. 

The Supraorbital Notch and its Relation to the Eyebrow Eminence. 

Schwalbe has pointed out that the supraorbital notch bears an important relation 
to the eyebrow eminence. A glance at figs. 17, 18, and 19, PI. I., which exhibit the 
region in the baboon, shows that this notch is the starting-point on the margin of the 
orbital opening from which the oblique groove, which intervenes between the super- 
ciliary and supraorbital elements, proceeds. Even in Type III. of this region, where the 
different elements are massed together with no external indication of their separate 
identity, Schwalbe takes the supraorbital notch as giving the only clue to the de- 
markation of the superciliary and supraorbital elements of the eyebrow projection. 
Schwalbe has failed to appreciate, however, that the supraorbital notches in man 
and the lower apes are not morphologically equivalent, nor yet similarly placed on the 
margin of the orbital opening. In other words, the disposition of the frontal nerve in 
man and the ape is different. 

In man the frontal branch of the ophthalmic division of the trigeminal nerve 
pursues a straight course within the orbit upon the upper surface of the levator 
palpebrse superioris, and about midway between the inner and outer walls of the 
cavity. At a variable point it gives off its supratrochlear branch and is continued 
onwards as the supraorbital nerve. The supratrochlear nerve inclines inwards towards 
the inner wall of the orbit, and finally turns round the inner part of the orbital margin 
above the trochlea of the superior oblique muscle to gain the forehead. As a rule it 
leaves no mark upon the bone as it winds on to the forehead ; sometimes, however, its 
path is indicated by a groove, as in the New South Wales cranium (figs. 20 and 21, 
PI. II.), and at other times it may pass through a foramen, as in the case of the 
Neanderthal cranium (fig. 3, p. 288, and fig. 6, p. 298). These markings are more 
frequently present in the crania of lower races, and more especially in those with a 
projecting glabellar and eyebrow region. 

The supraorbital nerve, or the continuation of the frontal trunk, reaches the forehead 
by turning upwards in the supraorbital notch or foramen. This notch is variable in 
position, but usually it lies a little to the inner side of the mid-point of the supraorbital 

Mr Ninian Bruce, B.Sc, has kindly made dissections for me of the orbital cavity 
in three chimpanzees, one orang, one yellow baboon, and in several species of the genus 
Macacus. These have shown that the frontal nerve in the ape does not present the 
same relations within the orbital cavity as is the case in man. 

In the baboon and the macaque the frontal nerve does not divide into two branches 
within the orbit, but issues from that cavity in the form of one undivided trunk. 
Further, this nerve courses through the orbit in close relationship to its inner wall, and 


turns round a wide and conspicuous notch on the inner part of the supraorbital margin 
and immediately to the outer side of the glabella (fig. 19, PL I.). It follows from this 
that the supratrochlear nerve in man is the representative of the whole nerve in the 
baboon and macaque, and that the occasional groove or foramen on the orbital margin 
associated with it is the morphological equivalent of the large single notch in the lower 
ape. The new position of the frontal nerve in man (on the upper surface of the levator 
palpebrse superioris and midway between the outer and inner walls of the orbit) and 
the passage of the greater number of its fibres through a new nerve (the supraorbital) 
is a condition which has probably been brought about by the increase in the breadth of 
the human forehead, which renders it necessary for the greater number of the group of 
nerve-fibres which go to supply the skin of this area to be shifted in an outward 
direction. Like most acquisitions of recent phylogenetic origin, the condition is one 
which is liable to considerable variation. Not only is the relative size of the two 
branches of the frontal nerve in man subject to variation, but also the position of the 
supraorbital notch on the orbital margin is very far from being constant. 

The arrangement of the frontal nerve in the orbit of the chimpanzee and orang is 
slightly different from that which was seen in the baboon and macaque. The nerve still 
clings to the inner wall of the orbit, but when it comes to the region above the trochlea 
it divides into two branches, which no doubt correspond to the supratrochlear and 
supraorbital branches in man. This division takes place below the front part of the 
torus, and the supratrochlear is carried onwards in an upward and inward direction 
around the inner part of the supraorbital arch and immediately to the outer side of the 
glabella. The outer branch or the supraorbital nerve turns sharply outwards on the 
under surface of the torus, and then winds on to the forehead in an oblique and often 
very obscure groove, which is the representative of the supraorbital notch in man. 

This arrangement of the nerve can usually be made out in a very distinct manner 
in the skull of the adult gorilla. From the markings on the bone it becomes evident 
that the main portion of the nerve turns over the torus close to the inner wall of the 
orbit in a very shallow groove which partakes more of the nature of a smooth pathway. 
This lies above and in front of the trochlear pit. 

In nine out of eleven gorilla skulls there was evidence that the frontal nerve had 
divided close to the margin of the orbital opening, and further, that the outer branch 
(i.e. supraorbital) had diverged from the inner branch almost at a right angle before 
turning round the supraorbital margin. The two shallow pathways for these nerves 
are distinctly marked on the bone, and in a large male skull they were separated at the 
points at which they turned upwards by an interval of 10 mm. As a rule they are 
separated from each other by a low spine-like projection upon the under surface of 
the torus. 

In no sense either in the chimpanzee or the gorilla can either of these grooves be 
taken as giving an indication of the line of fusion between the superciliary and supra- 
orbital elements of the torus supraorbitalis. This can be seen in those young specimens 




of chimpanzee skulls to which we have referred as showing in more or less distinct out- 
line the superciliary element as a part of the torus. 

The foregoing facts are important in connection with Schwalbe's views in regard 
to the supraorbital notch, and they are also of high interest when we apply them to 
the information we possess regarding the eyebrow region in the Neanderthal cranium. 

Schwalbe has given us a most careful account of the orbital roof and supraorbital 
margin in the Neanderthal cranium (8), and has supplemented his description by an 
instructive outline drawing (fig. 6, p. 298) and a photograph (fig. 7, PI. I.) of the 
under surface of the frontal bone. From these it is apparent that to a large extent 
the human arrangement of the frontal nerve has been present in the Neanderthal race. 
The division of the nerve-trunk is clearly indicated by groovings on the orbital roof, 
and the presence of a foramen for the supratrochlear part and a notch for the supra- 

Fig. 13. — Diagram to show position of frontal nerve and its branches in the 

Neanderthal specimen and in recent man. 
a. Supraorbital nerve (recent man). b. Supratrochlear nerve (recent man). 
a',b'.c'. Markings on roof of orbit of c. Frontal nerve (recent man). 
Neanderthal cranium. d. Supratrochlear foramen. 

orbital part are clearly delineated — more especially on the right side. But whilst this 
is the case, certain pithecoid characters are evident : ( 1 ) the groove on the orbital roof 
which conveyed the frontal nerve lies nearer to the inner wall than we are in the 
habit of seeing it in recent man ; (2) the two branches of the nerve diverge widely 
and abruptly from each other like the limbs of the letter Y, and of the two the supra- 
trochlear branch seems to lie more in the line of the parent trunk than the supraorbital ; 
(3) the groove and the foramen for the supratrochlear branch show that this nerve 
must have been unusually large, i.e. judged from the human standpoint. These 
features bespeak a nearer approach to the ape ; but in making this statement we must 
not lose sight of the fact that in several Australian crania in the Museum of the 
University there are markings which also point to pithecoid leanings of a some- 
what similar nature. 

But the interest in the nerve-markings is not exhausted by a study of the supra- 


trochlear and supraorbital grooves, notches, and occasional foramina. Anyone who 
makes a comparative study of the region will be naturally led to inquire, in the case of 
Type II., whether the sulcus which separates the superciliary and the supraorbital elements 
is not in great part produced by the supraorbital nerve in the first part of its course 
on the forehead. When I began my investigation this view appealed strongly to me, and 
gained force from a dissection which I made of a young yellow baboon, in which the 
nerve undoubtedly occupied the whole length of the sulcus in question. Subsequent 
research, however, has shown that, while the nerve may turn outwards in the groove for 
a short distance, the sulcus cannot in any sense be regarded as a nerve pathway. 

Having now taken a general survey of the morphological characters of the eyebrow 
eminences in man and the apes, we are in a position to discuss the view expressed by 
Professor Schwalbe that these features alone are sufficient to constitute a specific 
difference between the Neanderthal race and all other races of mankind. In coming to 
this conclusion, it seems to us that Schwalbe in some degree exaggerates the importance 
of the features in question, and has failed to take into account the numerous individual 
crania of recent races which are found to some extent filling up the gap which exists 
in this respect between the Neanderthal cranium and the crania typical of the races of 
the present day. 

The leading peculiarities of the Neanderthal eyebrow region are : — 

1. Its striking degree of prominence. 

2. Its external configuration and its morphological constitution. 

3. Its relation to the cerebral part of the frontal bone and to the roof of the orbit. 
In so far as 1 and 2 are concerned, the Neanderthal cranium may be said to fall within 
the limits of the races which exist at the present day. As we have seen, individual 
crania are found amongst the Australians which exhibit very nearly, if not quite, as 
great a prominence in the glabellar region. Nor are the general configuration and the 
constitution of the eyebrow projection which are characteristic of the Neanderthal group 
peculiarities which belong to this race alone. In these respects they are associated in 
Type III. with many individuals of other existing races. 

As Schwalbe has recognised, it is in the relation of the supraorbital projection to the 
cerebral and the orbital cavities that the most important distinction lies. But even 
in this respect we find in recent man transition forms, and we have indicated the New 
South Wales cranium as a case in point. Further, it has long been known that in the 
Australian, notwithstanding the projection of the eyebrow region, the tendency is 
towards the formation of frontal air-sinuses of small size. In the Australian skull, as in 
the Neanderthal cranium, the eyebrow eminence is as a rule largely formed by a massing 
of the bone in front of the air-chamber. 

It is doubtful if much proKt can be derived from a discussion as to whether the 


eyebrow characters in the Neanderthal cranium are to be accorded a specific value or 
not. A question of this kind is decided on more or less conventional grounds, and must 
be answered by each one for himself according to the interpretation which he puts upon 
the word " specific." At the same time it must be admitted that if we examine the 
basis on which zoological classification rests we shall find many cases in which species 
are determined upon evidence more slender than that which Schwalbk brings forward 
in favour of establishing the species of Homo primigenius for the reception of the 
Neanderthal people. 

There cannot be a doubt that the formation of such a species would be convenient 
in many ways ; but even allowing for the vague and indefinite understanding which exists 
amongst biologists regarding the determination of what characters should be elevated 
to the plane of specific importance and what characters should not, I can hardly bring 
myself to believe that we would be justified in adopting this course from the characters 
presented by the supraorbital region in the Neanderthal race. 


(1) Bianchi, "J seni frontali e le arcate sopraciliari studiate nei crani del delinquent!, degli alienati e dei 

normal!," Archivio per Vantrop. e /,'einologia, 1892. 

(2) Cunningham, D. J., " The Head of an Aboriginal Australian," Journ. Royal Anthrop. InstiL, vol. xxxvii. 


(3) ,, The Australian Forehead : Anthropological Essays presented to Edward Burnett 

Tylor, in honour of his 75th birthday, 2nd October 1907, Oxfurd, 
Clarendon Press. 

(4) Giufrida-Ruggeri, Dr V., " Nuove ricerche morfologiche e craniometriche," Atti della Societal Roman a 

di Antropologia, vol. viii., fasc. 1, 1901, p. 21. 

(5) Schwalbe, G., " Uber die Schadelformen der iiltesten Menschenrassen rait besonderer Beriicksichtigung 

des Schadels von Egisheim," Mitieil. d. philomath. Gesellschaft in Elsass- 
Lothringen, v., 1897. 

(6) ,, " Studien liber Pithecanthropus erectus (Dubois)," Zeitschrift fur Morphologie und 

Anthropologic, Band i., 1899. 

(7) ,, " Uber die iiltesten Menschenrassen," Wiener Minische Wochenschrift, 1899, No. 1. 

(8) ,, " Der Neanderthalschiidel," Bonner Jahrhilcher, Heft 106, 1901. 

(9) „ " Zur Frage der Abstammung des Menschen," Zeitschrift fur Morphologie und Anthropo- 

logic, Mai 1906. 

(10) ,, " Das Schadelfragment von Briix und verwandte Schadelformen," Zeitschrift fur Morpho- 

logie und Anlhropologie, Mai 1906 (Sonderheft). 

(11) ,, "Das Schadelfragment von Canstatt," Zeitschrift fur Morphologie und Anlhropologie, 

Mai 1906 (Sonderheft). 

(12) Selenka, Dr Emil, Menschenaffen : Studien uber Eidwickelung and Schadelbau, zweite Lieferung, 11 

Kapitel : "Schadel des Gorilla und Schimpanse," 1899. 

(13) SOLLAS, W. I., "On the Cranial and Facial Characters of the Neanderthal Race," Phil. Trans., series B, 

vol. cxcix. (pp. 281-339). 

(14) Turner, Logan, The Accessory Sinuses of the Nose, Kdin., 1901. 

(15) Zuckkijkandi,, E., Normole und pathologische Anatomie der Nasenhuhle und Hirer pneumatischen 

Anhdnge, 1. Band, 2. Auflage, 1893. 



Plate I. 

Fig. 14. Skull of a young Macaeus cy clops. The superciliary eminences are seen in the form of two 
isolated, oval, slightly raised areas. (Type I.) 

Fig. 15. Skull of Macaeus rhesus. Superciliary eminence strongly marked, elongated, and separated 
from the supraorbital margin in its whole length. (Type I.) 

Fig. 16. Skull of adult chimpanzee (Goodsir series). Torus supraorbitalis well marked, but it will be 
noticed that the superciliary eminence has not been completely absorbed into it. The general outline of the 
superciliary eminence is distinct, and it also presents a distinctive texture. (Type II. merging into Type III.) 

Fig. 17. Skull of a young chacma baboon (Cynocephalus porcarius, Turner series). The superciliary 
eminences stand clear of the supraorbital margin. (Type I.) 

Fig. 18. Skull of an adult baboon (species doubtful). The superciliary eminences by their inner parts 
are confluent with the inner parts of the supraorbital margins. (Type II.) 

Fig. 19. Skull of an adult Anubis baboon {Cynocephalus anubis). Shows a nearer approach to the fusion 
of the superciliary and supraorbital elements of the region than in fig. 18. 

Plate II. 

Fig. 20. New South Wales cranium (xxix. B. 1, Turner series). Tori supraorbital es well marked. 
(Type III.) a, supratrochlear grooves ; b, supraorbital notches. 

Fig. 21. New South Wales cranium (xxix. B. 1). Orbital aspect of the frontal bone, a, supratrochlear 
grooves ; b, supraorbital notches. Note how the interorbital region is pinched in between the two supra- 
trochlear nerves. 

Fig. 22. Australian skull from the Queensland district (xxix. A. 10, Turner series). Tori supraorbitales 
with a steep and highly curved forehead. (Type III.) 

Plate III. 

Fig. 23. Skull of a native of New Hebrides Islands (xxvii. C. 3, Turner series). Type II. of the supra- 
orbital region. Contrast with fig. 18, PI. I. 

Fig. 24. Skull of native of Australia from Riverina district (xxix. B. 12, Turner series). Very pro- 
jecting supraorbital region ; transition condition between Type II. and Type III. 


z^ty, Jtim*y\ 

Fig. 14. 

Fig. 17. 

Fig. 15. 

Fig. 16. 

Fig. 18. 

Fig. 19. 

Trans. Roy. tioc. Edin. Vol. XLVI. 

Professor ]J. J. Cunningham on " TI13 Evolution of the Eyebrow Region of the Forehead." — Plate II. 

Trans. Roy. Soc. Edin. Vol. XLVI. 

Professor I>. J. Cunningham on "The Evolution of the Eyebrow Region of the Forehead." — Plate III. 

Groove along line of , 
union between super- \ 
ciliary and supraorbi- \ 
tal elements. 

Fig. 23. 


Fig. 24. 


( 313 ) 

XIII. — On the Age of the Old Red Sandstone of Shetland. By John S. Flett, 

M.A., D.Sc. 

(Read March 18, 1901. MS. received May 26, 1908. Issued separately July 8, 1908.) 

In spifce of its remote situation, the Old Red Sandstone of Shetland attracted a con- 
siderable amount of attention from geologists during the last century. It is exposed in 
excellent coast sections, which often yield very beautiful cliff scenery ; and, in addition 
to being the most northerly of the stratified rocks of Great Britain, it includes a rich 
succession of volcanic and intrusive rocks which are of great interest and variety. The 
axis or backbone of the Shetland archipelago consists of gneiss, mica schist, slate, and lime- 
stone, with epidiorites, serpentine, and talc schists. On each side of this there is an area 
of Old Red Sandstone ; that on the east extending from Sumburgh Head, in the extreme 
south, to Rovey Head, a little north of Lerwick, and comprising also the islands of Bressay, 
Noss, and Mousa. On the west side of Shetland the Old Red Sandstone Series is much 
altered, probably by the heat of the granite and other intrusive rocks, so that they often 
have the appearance of quartzite, and were for a long time regarded as belonging to the 
metamorphic series. In 1879, however, Peach and Horne (28) showed that, in places, 
they contained fossil plants which indicated that they belonged to the Old Red Sand- 
stone formation. 

The earliest accounts of the Old Red rocks of Shetland are to be found in the 
descriptive works of Jameson (1G), Neill (25), Boue (1), Shireff (34), Fleming (2), 
and Hibbert (14). These writers were all of the Wernerian school, and described the 
conglomerates, sandstones, and flags as " secondary," resting on the " primitive " or 
metamorphic group. Of these accounts the best are those of Hibbert and of Fleming ; 
the latter in particular deserves mention, as being the first to record the occurrence of 
fossil plants in the Lerwick Sandstones. 

In 1853 an important advance was made by the description of some fossil plants 
from South Ness, Lerwick, by Dr (afterwards Sir) Joseph D. Hooker (15). He referred 
them to two species of Calamites. This paper was communicated to the Geological 
Society of London, and was accompanied by a note by Sir Roderick Murchison (21), in 
which he stated his conviction that " the sandstone of Lerwick is of the same age as the 
rocks of Elgin, Burghead, Tarbet Ness in Ross, and Dunnet Head in Caithness, all of 
which Professor Sedgwick and myself described as constituting the uppermost member 
of the Old Red Sandstone, and as overlying the Caithness flagstones, with their numerous 

Murchison, accompanied by Sedgwick, had already visited Caithness, Ross, and 
Cromarty, and was familiar with the Old Red Sandstone of these districts (33). He subse- 
quently proceeded again to Caithness, and thence to Orkney and Shetland (5). The 



impressions he received on this visit confirmed the opinion he had already formed, and 
led him to place these beds definitely in the younger portion of the Old Red Sandstone 
(22), (23). 

The subdivision of the Old Red Sandstone of Scotland into Lower, Middle, and Upper, 
which Sir R. Mdrchison had advocated, was discarded by Sir Archibald Geikie (6) in 
his well-known paper on the Old Red Sandstone of Western Europe (Part I.), which still 
forms the principal source of information regarding the Orcadian Old Red Sandstone. 
He proved definitely that the Upper rested in Hoy and elsewhere, with a marked 
unconformability, upon an eroded surface of the Orcadian (7) ; but the Middle and Lower 
subdivisions of Murchison he grouped into one. The evidence of the fossil fishes and 
fossil plants points to their being distinct formations; and in its recent Memoirs (19) 
the Geological Survey of Scotland has reverted to the threefold grouping of Murchison. 

In his paper, Sir A. Geikie does not express any decided opinion as regards the exact 
horizon of the Shetland Old Red. He recognises that Murchison had relied mainly on 
Hooker's determination of the Lerwick plants as Calamites in assigning these beds to 
the topmost portion of the system ; and as this identification had been shown to be 
dubious (32), the conclusion arrived at was hardly valid. The great resemblance of the 
volcanic rocks in this series in Shetland to those in the Caledonian (Lower) Old Red of 
Scotland, and the occurrence in Shetland of Estheria membranacea, known also in the 
flagstones of Caithness and Orkney, were pointed out, and no doubt led him in sub- 
sequent years to include the Shetland beds with his Lower Old Red Sandstone. At 
any rate, this is the correlation that was ultimately accepted by him (8), (9). 

About the same time as Sir Archibald Geikie's paper appeared, Professor Heddle 
published his geognosy of Shetland (13), in which brief space is given to the Old Red 
Sandstone. In Dr Gibson's account of the Old Red Sandstone of the East of Shetland 
(12) very careful descriptions of the lithology of the beds are given, but the lack of fossil 
fishes is deplored. In the absence of more definite evidence, it is assumed that the 
horizon of these beds is the same as that of the Caithness flags. 

In 1879 the first of a series of papers on the geology of Shetland by Dr Peach and Dr 
Horne appeared in the Quarterly Journal of the Geological Society of London (28). This 
included a description of the Old Red Sandstone, and was followed by two others. The 
remarkable series of volcanic rocks was specially investigated (29). It is not too much to 
say that, as the result of their work, amplifying and correcting the earlier descriptions of 
Hibbert, Heddle, Geikie, and Gibson, the geology of Shetland, varied and complex 
though it is, is better known than that of any part of Scotland which has not been 
mapped by the Geological Survey. Four excellent maps of the geology of Shetland 
have been published, one by Professor Heddle (13) and three by Dr Peach and Dr 
Horne (28), (29), (30). 

To their accounts of stratigraphy of the Old Red Sandstone of Shetland, and its 
relations to the older metamorphic rocks, little remained to be added. As regards the 
age of these beds they maintained a conservative attitude, though acquiescing in Sir 


A. Geikie's relegation of them to the Lower division of the system (which, of course, 
included the Orcadian). In this opinion they were supported by C. W. Peach (27), who 
re-examined the fossil plants described by Hooker, and found that they exhibited close 
affinities to those obtained in the Old Red of Caithness and Orkney. 

In 1898, after spending part of several years in investigating the Old Red Sandstone 
of Orkney (3), I determined to visit Shetland and make a search for fossil fishes which 
would establish the position of the sandstones and shales of these islands relatively to 
those of Caithness and Orkney. Six weeks were spent in a scrutiny of all the best ex- 
posures on the east side of the mainland from Sumburgh Head to Lerwick, and in the 
islands of Bressay and Noss. The results, though unsatisfactory, were not entirely dis- 
appointing, as indecipherable fragments of fishes were obtained in Bressay, at Lerwick, 
Sandwick, and the east side of Quendale Bay. Further search was determined on ; and 
to meet the expenses of quarrying, a grant was applied for from the Royal Society of 
London (Government Grant Committee), which was conceded. Consequently, in 1899, 
with the consent of the late Mr Hamilton, of Unst, an opening was made in the beach 
on the east side of Cullingsburgh Voe in Bressay, and our expectations were soon con- 
firmed by the discovery of scattered plates of undoubted fossil fishes belonging to new 
species. These fishes were handed to Dr Traquair for identification, and a preliminary 
notice was inserted in Nature to announce the discovery of a new zone of the Old Red 
Sandstone of Scotland (24). 

The Fish-bearing Beds of Bressay, and their Position in the Old Red 

Sandstone of Shetland. 

As recognised by all who have described this area, the structure of the district around 
Lerwick and Bressay is exceedingly simple. A little west of Lerwick, coarse conglomerates 
are faulted against the metamorphic series. They dip towards the east, and are succeeded 
at the town of Lerwick by reddish and grey sandstones, often current-bedded, and 
sometimes containing large rounded pebbles of quartzite, granite, etc. At the point 
south-east of Lerwick known as the Nabb, grey micaceous sandstones occur, full of plant- 
remains, and containing also the small crustacean Estheria membrancea (22). On the 
opposite shore of Bressay Island the first beds met with are brownish and grey sand- 
stones, often conglomeratic, and sometimes brecciform, with occasional grey and reddish 
shales. A series of faults or crush belts run nearly north and south along this side of 
the island from Maryfield to Ham, setting the beds frequently on end, and converting 
them into breccias and crush conglomerates. In crossing Bressay the dip of the rocks 
is consistently east or south-east, varying from ten to thirty degrees. The commonest 
rocks are grey, micaceous, thin-bedded sandstones, with coarser, less micaceous, gritty 
seams, often current-bedded. The sandstones contain rounded clay galls, and their sur- 
faces are often covered with blackened fragments of plants and shreds of fine shale. Small 
faults are frequently seen in the coast sections, mostly running parallel to the strike. On 


the west side of Noss Sound a small volcanic neck, described by Peach and Horne (29), 
occurs, with a thin bed of ash. Faulting also is visible here, and the beds are often 
steeply inclined. In Noss the easterly dip again prevails, and in the great cliff on the 
east side of this island a fine section of thin flaggy sandstone and grey shales is 
exposed, exceedingly well stratified, and resembling closely many of the cliff features of 
Orkney and Caithness. 

In view of the persistent easterly dip, often at fairly high angles, the whole thick- 
ness of this series must be several thousand feet ; but the evidence of faulting along the 
shores of the Sounds is sufficient to render exact estimates impossible. The fish beds 
in Cullingsburgh Voe are rather above the middle of the Bressay Sandstones. The 
fossils occur in a thin-bedded, flaggy, grey micaceous sandstone, and the plates are 
black in colour and well preserved. With them thin black impressions of plants are 
exceedingly common. The strata were evidently laid down in shallow water, close to 
land ; and the general facies of the rocks and of the fauna is in harmony with the 
supposition that they were fresh-water deposits. 

Fossiliferous bands must occur in Bressay besides that in which our excavations 
were made, as we found a fish fragment in a beach stone on the east side of Cullingsburgh 
Voe, and another in a loose rock to the west of the houses of Cullingsburgh. Professor 
Heddle (20), (10) tells us that he saw " specimens of small fishes, apparently acanthoides, 
embedded in a fine-grained muddy sandstone ; they were stated to occur in a quarry north 
of Gardie, in Bressay." There is no reason to doubt this record ; but, unfortunately, the 
fishes can no longer be traced (if they were preserved). I made a careful search in all 
the quarries near Gardie, but could see no remains of fossil fishes. It is clear that the 
Shetland Old Red Sandstone is by no means so barren as has hitherto been supposed ; 
but to obtain good fossils, great skill and patience, with some measure of luck, will be 

The Fauna op the Bressay Sandstones. 

The fish-remains obtained in these beds have been determined by Dr Traquair to 
belong to Asterolepis (sp.) and Holonema (sp. nov.) ; possibly there are also fragments 
which may be referred to a species of Holoptychius. Of these genera, the first is 
typically Upper Old Red ; it occurs also in beds assigned to that series at Nairn. 
Holonema is a genus founded by Newberry (26) for remains from the Chemung beds 
(Upper Devonian) of North America. Holoptychius is a very characteristic Upper Old 
Red genus. So far, then, the evidence of the fish fauna points clearly to the Shetland 
beds belonging to the Upper part of the Old Red system. It is not, however, absolutely 
satisfactory when closely examined. The Holonema is a new species ; the Asterolepis 
also is probably new, though as yet not definitely named ; and caution is necessary in 
classifying Old Red faunas in such cases. As an example of this we may quote the 
typical Lower Old Red genus Cejihalaspis, which occurs in the Middle Old Red of 


Caithness (35) and the Upper Old Red of Canada (36). Coccosteus also is known from the 
whole range of the Devonian or Old Red Sandstone succession (19). It is easy, however, 
to exaggerate the importance of exceptional cases. More important is the fact that 
Asterolepis is represented by a species as yet undescribed in beds immediately under- 
lying the John o' Groats zone of the Orcadian Old Red in Orkney (4). This indicates 
that if the Shetland beds be Upper Old Red, they have close relationships to the 
Orcadian. Similarly, there are grounds for believing that the Nairn beds (18) which 
contain Asterolepis are older than the Elgin, Dura Den, Jedburgh, and Kiltorcan 
Sandstones, in which Bothriolepis and Holoptychius nobilissimvs occur. The latter 
must represent the highest portion of the Old Red Sandstone, as in places they pass 
up conformably into the base of the Carboniferous. 

The only other animal remains as yet obtained from the Shetland Old Red Sandstone 
are Estheria membranacea (Pacht) and plant fragments. Of these, the former was first 
recorded by Sir R. Mukchison (22), and was described by Professor Rupert Jones (31). 
This species is abundant and well preserved in the Orcadian Old Red of Caithness and 
Orkney. It occurs also in the Devonian of Livland (Livonia) in Russia, but has not 
been found in the Upper Old Red Sandstone of any part of Great Britain (except 
Shetland). This is sufficient to indicate that the Shetland beds have close relationships 
with the Middle Old Red Sandstone. 

The Fossil Flora of the Shetland Old Red Sandstone. 

The fossil plants found near Lerwick have been described by many palseobotanists. 
They are in a most unsatisfactory state of preservation, but (with the exception of Sir 
Joseph Hooker) (15) all who have examined them have noted their resemblance to the 
hardly less imperfect plant-remains so numerous in the Middle Old Red of Caithness and 
Orkney. Recognising that little was definitely known about their real nature, I had 
specimens forwarded to Dr Kidston, in order to obtain the latest and most authoritative 
opinion regarding them. He assures me that, whatever may be their botanical affinities, 
they have nothing in common with the Lower Old Red flora of Forfarshire and Perth- 
shire, or the Upper Old Red floras of Kiltorcan and Roxburghshire. So far as they are 
determinable, they resemble rather the plants of the Orcadian Old Red. This opinion, 
though vague, is of value, as confirming the evidence provided by the crustaceans 
(Estheria) and one of the genera of fishes (Asterolepis). 


The lithological similarity between the Old Red Sandstone of Shetland and that of 
Caithness and Orkney has been remarked on by various writers. It is best seen in the 
cliff exposures when viewed from a distance, as the thin-bedded flags of Bressay and 
Noss, where eroded by the sea, yield cliffs very like those of Orkney. Closer examination, 


however, hardly strengthens the comparison. The Shetland beds are coarser, more 
micaceous, and less uniform than those of Orkney. Clearly they were formed by the 
denudation of a slaty and gneissose series, like that which forms the axial ridge of 
Shetland. Evidence of this sort does not carry much weight ; but we may note that 
the Old Red Sandstone of Shetland certainly resembles the Caithness beds in litho- 
logical character more closely than the Upper Old Red Sandstone of Hoy, Ross-shire, 
and the districts south of the Moray Firth. 


In speculating on the age of the Shetland beds, the character of the vertebrate fauna 
is by far the most reliable index ; and next to it we may place the evidence of the 
plants. The former points unmistakably to Upper Old Red conditions, the latter to 
close affinities with the Orcadian. The conclusions which at the present time we are 
justified in arriving at may be summarised as follows : — 

(1) The Old Red Sandstone beds of Shetland belong to a distinct zone of that 
system, different from any other yet identified in Britain. 

(2) By the evidence of its fossil fishes, it is most naturally placed in the Upper 

(3) Both the fauna and the flora indicate affinities with the Orcadian or Middle Old 

Red of Caithness and Orkney, consequently it may be the lowest or one of the lowest 
zones of the Upper Old Red Sandstone. 



(1) Bouk, Ami, Essai Geologique sur I'Ecosse, p. 114 (n.d. probably 1820). 

(2) Fleming, J., Mem. Wernerian Soc, vol. i., p. 174 (1808). 
Flett, John S., " On the Discovery in Orkney of the John o' Groat's Horizon of the Old Red Sandstone," 

Proc. Roy. Phys. Soc. Edin., vol. xiii., p. 255 (1896). 
„ „ " The Old Red Sandstone of the Orkneys," Trans. Roy. Soc. Edin., vol. xxxix., p. 383 

Ibid., p. 409 (1898). 

Memoir of Sir R. Murchison, vol. i., p. 139 (1875). 
"The Old Red Sandstone of Western Europe (Part I.)," Trans. Roy. Soc. Edin., 

vol. xxviii., p. 345 (1878). 
"The Old Man of Hoy," Geol. Mag., p. 49 (1878). 
Geological Sketches, p. 76 (1882). 
Text-Book of Geology, p. 1010 (1903). 

Geological Map of Scotland (published by Bartholomew, Edinburgh). 
Tram. Roy. Soc. Edin., vol. xxviii., p. 418 (1878). 
Text- Book of Geology, p. 1012 (1903). 
Gibson, George, Hie Old. Red Sandstone of Shetland (1877). 


>• »> 


Geikik, Sir A 


M )> 


)> )> 

>i i) 


• j j) 


)) >> 


Jl »» 


)) j) 


Gibson, Geor 

(16) { 


(13) Hbddle, M. F., Mineralogical Magazine, vol. ii. (1879). 

(14) Hibbert, Samuel, A Description of the Shetland Islands, p. 160, etc. (1822). 

(15) Hooker, Joseph D., "Note on the Fossil Plants from the Shetlands," Quart. Jour. Geol. Sac., vol. ix., 

p. 49 (1853). 

Jameson, R., Outline of the Mineralogy of the Shetland Isles, p. 15, etc. (1798). 

,, ,, The Mineralogy of the Scottish Isles, vol. ii., p. 195, etc. (1800 and 1813). 

(17) Kidston, E., Catalogue of Palaeozoic Plants in the British Museum, pp. 233 and 238 (1886). 

(18) Mackie, W. S., Sand Grains, Elgin (1894). 

(19) Memoirs of the Geological Survey of Scotland : " The Geology of Lower Strathspey." 

(20) Min. Mag., vol. ii., p. 156. 

(21) Murchison, Sir Roderick I., "Note on the Age and Relative Position of the Sandstone containing 

Fossil Plants in the Shetland Isles," Quart. Jour. Geol. Soc, 
vol. ix., pp. 50 and 50 bis. (1853). 

(22) ,, ,, " On the Succession of the Older Rocks in the Northernmost Counties 

of Scotland, with some Observations on the Orkney and Shetland 
Islands," Quart. Jour. Geol. Soc, vol. xv., p. 353 (1859). 

(23) „ „ Siluria (4th ed.), p. 259(1867). 

(24) Nature, Nov. 23, 1899, vol. lxi., p. 80. 

(25) Neill, Patrick, A Tour through some of the Islands of Orkney and Shetland, p. 72, etc. (1806). 

(26) Nbwberry, J. S., Palosozoic Fishes of N. America, p. 92 (1889). 

(27) Peach, C. W., Quart. Jour. Geol. Soc, vol. xxxv., p. 811 (1879). 

(28) Peach, B. N., and Horne, J., " The Glaciation of the Shetland Isles," Quart. Jour. Geol. Soc, vol. 

xxxv., p. 778 (1879). 

(29) „ „ " The Old Red Volcanic Rocks of Shetland," Trans. Roy. Soc. Edin., vol. xxxii., p 359 


(30) „ „ "The Geology of Shetland," in Tudor's The Orkneys and Shetland { 1883). 
, Rupert Jones, T., Quart. Jour. Geol. Soc, vol. xv., p. 404. 

) „ ,, On Fossil Estherise and their Distribution, ibid., vol. xix., p. 140 (1862). 

1 ,, ). "Monograph of the Fossil Estherise," Palseontographical Society, p. 14 (1862). 

Geol, Mag., 1890, pi. xii., and 1891, p. 50. 

(32) Salter, J. W., " On some Remains of Terrestrial Plants in the Old Red Sandstone of Caithness," Quart. 

Jour. Geol. Soc, vol. xiv., p. 72 (1858). 

(33) Sedgwick, A., and Murchison, E. I., "On the Structure and Relations of the Deposits contained 

between the Primary Rocks and the Oolitic Series in the North of Scotland," 
Trans. Geol. Soc, ser. ii., voL iii., p. 125 (1828). 

(34) Shireff, Agriculture of the Shetland Islands (1814). 

(35) Traquair, R. H., "On Cephalaspis magnifica (Traq.)," Proc Roy. Phys. Soc. Edin., vol. xii., p. 269 


(36) ,, ,, "Notes on the Devonian Fishes of Scaumenac Bay and Campbelltown in Canada," 

Geol. Mag. [3], vol. vii., 1890, p. 16. 

(37) Tuffnell, Henry, "Notice of the Discovery of Fossil Plants in the Shetland Isles," Quart. Jour. Geol. 

Soc, vol. ix., p. 49 (1853). 

( 321 ) 

XIV.— On Fossil Fish-remains collected by J. S. Flett, M.A., D.Sc, from the Old 
Red Sandstone of Shetland. By Ramsay H. Traquair, M.D., LL.D., F.R.S. 
(With Two Plates.) 

(Read March 18, 1901. MS. received May 8, 1908. Issued separately July 4, 1908.) 

Little has hitherto been known about the animal remains of the Old Red Sandstone 
of Shetland. In 1858 Sir Roderick Murchison mentioned the occurrence, in flaggy beds 
in the environs of Lerwick, of " the same little Crustacean (the Estheria) which occurs 
at Thurso and Kirkwall."* Heddle in 1878 refers to the occurrence of fish-remains in 
the rocks of the same region in the following terms : — -" Specimens of small fishes, 
apparently acanthoides, were shown the writer ; these were imbedded in a brown 
fine-grained muddy sandstone ; they were stated to occur in a quarry north of Gardie 
in Bressay." f Again, in the same year, Sir A. Geikie, in his well-known paper on the 
" Old Red Sandstone of Western Europe," states that Dr Heddle " informs me that he 
was shown some ichthyolites (Coccosteus, etc.) in Bressay, which he was assured had 
been found among the flagstones of that island. "| Unfortunately it does not appear 
that any description of these specimens has ever been published, nor, so far as I am 
aware, is their present whereabouts known. 

In 1898, however, Dr Flett obtained from the flagstones at Bressay some 
fragments of undoubted dermal plates of fishes. In the following summer (1899), 
aided by a grant from the Royal Society of London, the same investigator succeeded 
in collecting from the same beds a larger number of these remains, which their finder 
was so kind as to refer to me for study and description. These specimens, though still 
very fragmentary, were nevertheless good enough, some of them, to afford a clue to 
their genera ; and accordingly, in a brief notice which appeared in Nature of 
November 23 of that year, it was announced that fragments referable to Holonema, 
Newberry, and Asterolepis, Eichwald, had been identified by me among Dr Flett's 

This was of great geological interest, as the first named genus has as yet been only 
with certainty found in strata of Upper Devonian age, namely, in the " Chemung " 
rocks of North America. Asterolepis, on the other hand, though well known in the 
Upper Old Red of Scotland and Russia, is also not unknown in the Orcadian series 
(Middle Old Red of Murchison), Dr Flett and Mr Spence having collected speci- 

* Quar. Journ. Geol. Soc, vol. xv., 1858, p. 413. 

t "On the Geognosy of Scotland. The Mainland of Shetland," Mineralog. Mag., vol. ii., No. 11, Dec. 1878, 
p. 156, footnote. 

J Trans. Roy. Soc. Edin., vol. xxviii., 1878, p. 418. 



mens referable to the genus both from the Mainland of Orkney and from South 
Ron aid shay. 

Discussion of the geological position of the Old .Red of Shetland I must, however, 
leave to Dr Flett, and so I proceed to my own proper task, namely, the description 
of the fish-remains themselves. 

Class PISCES. 



Family Asterolepid^e. 

Genus Asterolepis, Eichwald. 

Asterolepis, sp. indet. PI. I. figs. 1-4. 

Description. — Owing to the fragmentary and otherwise badly preserved condition 
of the remains referable to Asterolepis, it is impossible to give a definition of the species 
to which they belong, and consequently I refrain from giving it a name at present. 

Fig. 1.— Outline of impression of the greater part of left posterior dorso-lateral plate of Asterolepis, the missing parts 
given in dotted lines. A anterior, 13 posterior extremity ; a.a., area overlapped by anterior median dorsal plate : b.b., area 
overlapped by posterior median dorsal plate ; c, part of the area overlapped by the anterior dorso-lateral. The lower part of 
the plate with the arese overlapped respectively by the anterior and posterior ventro-lateral plates is deficient, and is represented 
here by dotted lines. Compare with fig. 1 on Plate I. 

Fig. 1, PI. I., is the impression of the outer surface of a plate which is at once 
identifiable as the greater portion of the left posterior dorso-lateral of a fish belonging 
to the genus Asterolepis. Its explanation is given in the text-figure above, in which 
A is the anterior and B the posterior end of the bone, the missing parts being restored 
in dotted outline. Now along the upper margin is the imprint of a narrow articular 
area aa, unfortunately broken away a little in the middle. This is the surface over- 
lapped by the anterior median dorsal plate. Then behind this area and along the 
posterior-superior margin is a second narrow articular area bb, namely, that overlapped 
by the posterior median dorsal. Herein we have the proof that the plate belongs to 
Asterolepis, as in this genus the anterior median dorsal overlaps, as above, the posterior 


dorsolateral, whereas in Pterichthys and Bothriolepis the reverse is the case, 
the anterior median dorsal being overlapped by, instead of overlapping, the plate in 

The impression of the exposed surface of the plate shows evidence of a tuberculated 
sculpture, but this is not sufficiently well preserved to enable us to compare it with 
that in any other species. 

Three other fragments are undoubtedly asterolepid, and may safely be referred to 
the same genus and species with the plate described above. 

Fig. 2, PI. I., is recognisable as the anterior part of the left anterior dorso-lateral 
plate, broken off behind in such a way that the fractured edge shows a backwardly 
directed point. Most of the bony matter has scaled off, so that the external sculpture 
is lost. The subjoined text-figure 2 gives the orientation of this fragment, the parts 
a wanting being, as in text-figure 1, indicated in dotted outline. Here the letter a 
indicates the marginal area overlapped by the anterior median dorsal plate, but the 
area along the opposite (lower) margin, overlapped by the anterior ventro-lateral, is 

Fig. 3 shows the distal segment of the pectoral limb, split up longitudinally, the 

Fig. 2. — Outline of a fragment of the left anterior dorso-lateral plate of Asterolepis, the contour being restored in dotted 
lines. A anterior, B posterior extremity ; a., area overlapped by the anterior median dorsal plate. 

outer surface being adherent to the matrix. The figure shows as much of its configura- 
tion as could be expressed in any description. 

In fig. 4 we have the articular element of the proximal segment of the pectoral 
limb. Most of its substance has been lost, and its contour also is not sufficiently 
perfect to enable us to determine whether it belongs to the right or left limb, or to the 
dorsal or ventral surface. 

Observations. — That the remains described above belong to the genus Asterolepis 
of Eichwald there cannot be the remotest doubt, and it is likewise probable that 
they represent a hitherto undescribed species. However, the scantiness of the material 
as well as its very fragmentary nature render it impossible to give a satisfactory specific 
diagnosis, and so, as already remarked, I refrain for the present from giving it a name. 

Locality. — Bressay, near Lerwick, Shetland ; collected by Dr J. S. Flett. 

* See the author's paper on the " Structure and Classification of the Asterolepidae " in Ann. and Mag. Nat. 
History [6], vol. ii. (1888), pp. 485-503. Also in " Fishes of the Old Red Sandstone," pt. ii. No. 1, Palceontographical 
Society, 1894. 


Incertse Sedis. 


Family (?) Coccosteid^. 

Genus Holonema, Newberry. 

Generic Characters. — Imperfectly known. The genus was instituted for thin, 
rather large plates covered externally by closely set ridges or rugse radiating from a 
central point towards the circumference, but often, too, especially in the middle of a 
plate, irregular or reticulated. The plates as yet known seem to belong to a ventral 
cuirass of Coccostean structure. 

History. — In 1883 the late Professor Claypole described and figured what he 
designated as a " large fish plate from the Upper Chemung (?) beds of North 
Pennsylvania."* This plate he considered to be probably the median ventral plate of 
a species of Pterichthys — the "lozenge" plate of Hugh Miller. 

The plate itself is of an oblong pentagonal form, having one very short side placed 
transversely at the supposed posterior extremity, followed by two pairs of laterally 
symmetrical sides, of which the shorter pair meet at an angle on the supposed anterior 
aspect. All round is " a flat, finely striated margin " — namely, the narrow area over- 
lapped by adjoining plates — within which the exposed surface is sculptured with 
radiating, closely set, subparallel rugse, more or less perpendicular to the margin, though 
in the centre of the plate tending to show an irregular reticulation. 

This plate, evidently a median one, Claypole considered to belong to a new species 
of Pterichthys, and to represent the median ventral or " lozenge " plate of Hugh Miller. 
Accordingly he named it Pterichthys (?) rugosus, noting, however, that Professor 
Newberry, to whom he had sent a photograph and description of the specimen, had 
informed him in a letter that he (Professor Newberry) " very much doubted if the 
plate here described belonged strictly to Pterichthys, and was inclined to consider it 
the type of a new genus." 

Six years afterwards Newberry, in his special work on the Palseozoic Fishes of 
North America,^ figured some further plates of the same fish from the "Chemung" 
strata, which he felt necessary to separate generically from Pterichthys, though retain- 
ing Claypole's name for the species, which he now designated Holonema rugosum. 
In PI. XVII. of the work he reproduces Claypole's figure (fig. 2) and adds representa- 
tions of three other pieces. In fig. 1 we have a rather large fragment, in which the 
rugse, frequently bifurcating, radiate from nearly a central point, while in fig. 4, 
obviously a fragment, the point of radiation is at one side ; lastly, in the remaining 
fragment, shown in fig. 3, the ruga3 are in some places contorted and irregular. It is 
to be noted that, although Newberry nowhere directly states his opinion as to the 

* Proc. American Phil. Soc, vol. xx. (1883), pp. 664-666, with figure. 

t Palaeozoic Fishes of North America (Washington, 1889), pp. 92-95, pi. xvii. figs. 1-4. 



systematic position of Holonema, he very significantly compares Claypole's original 
specimen with "the lozenge- shaped plate of Coccosteus; the ventro-median plate of Owen." 
Professor Claypole in the following year returned to the subject in a paper entitled 
" Palseontological Notes from Indianopolis."* Here he states that a second plate, 
evidently a lateral one, had come into his possession, and also that Professor H. S. 
Williams had obtained a new and unique specimen from the Catskill of New York, 
consisting of several large plates of Pterichthys rugosus, Claypole. He gives a restored 
outline of the cuirass, here reproduced, text-fig. 3, from which it will be seen that behind, 
or supposed to be behind, the original central plate is a posterior median one, while 
two pairs of lateral ones complete the sides. Quoting Professor Newberry's opinion 
that the characters justify a separation from Pterichthys, he says that, " for the present, 


Fig. 3. — Ventral cuirass of Holonema rugosum, supposed by Claypole to be the dorsal shield of a species of Pterichthys or 
allied genus. From Prof. Claypole's paper in the American Geologist quoted above. The nomenclature of the plates given by 
him is — a, dorso-median (first described) ; c, dorso-lateral left ; b x post-dorso-lateral right ; b 2 , post-dorso-lateral left ; d, 

According to Dr Smith Woodward's view, obviously the correct one! that this cuirass is the ventral shield of a Coccostean 
turned the wrong way, these plates would read— a, median ventral ; d, anterior median ventral ; b± and b 2 , anterior ventro- 
laterals ; c, posterior ventro-lateral. 

however, the matter must remain undetermined and the fossil be called Pterichthys 
(Bothriolepis) rugosus or Holonema rugosum.'" It is, however, clear that Claypole still 
believed in the Pterichthys theory, and also that he now looked on the plates in his 
figure as being dorsal instead of ventral ; for he says, " The form of the central plate, 
the presence of another median plate at its narrow end, the form, size, and fit of the 
lateral plates at its wide end, the direction of the wrinkles, and the number of plates 

* American Geologist, vol. vi. (1890), pp. 255-258. 


necessary to complete the shield (assuming its near relationship to Pterichthys and 
Bothriolepis, which seems reasonable) all combine to indicate the dorsal rather than the 
ventral aspect of the fish." 

But in 1891 Dr Smith Woodward, in the second part of his " Catalogue" (pp. 314- 
315), unhesitatingly classes the genus as Coccostean, pointing out the resemblance in 
form between the median plate figured by Claypole and Newberry and the median 
ventral of certain species of Coccosteus, e.g. C. disjectus ; holding also that "the recent 
description of the complete ventral shield by Claypole proves that it agrees with that 
of Coccosteus in every essential particular. The ' post-dorso-median ' plate of Claypole 
is obviously the anterior median ventral, while the ' post-dorso-lateral ' and ' dorso- 
lateral ' of the same author are the anterior and posterior ventro-lateral plates re- 
spectively." The cuirass figured by Claypole is in effect a Coccostean plastron turned 
the wrong way. 

In the same year Professor E. D. Cope # mentioned a specimen of the genus from 
Mansfield, Tioga County, Pennsylvania, namely, a lateral plate of the plastron. He 
also states : "Besides this there is a nearly complete pectoral spine, which is of much 

interest, as this part of the skeleton has not been previously known The spine 

differs from that of both Bothriolepis and Pterichthys in being without complete 
segmentation. It is continuous throughout to the apex. This then will constitute the 
generic distinction so far known between Holonema and Bothriolepis" Of this supposed 
spine of Holonema a figure is given, which at once betrays to the eye of anyone 
familiar with the structure of the Asterolepidse that it is not an entire appendage but 
merely the distal portion or segment of the pectoral limb of Bothriolepis. 

In another paper t published in 1892 Cope figured and described as Holonema 
rugosum a plate which he considered to belong to a new species. As regards the genus, 
he still looked upon it as belonging to the Antiarcha, but with tivo ventral median plates, 
thus differing from Bothriolepis. He also described without figure a plate which he 
referred to H. rugosum, saying regarding it : " This piece, together with the pectoral 
limb which I have already described, demonstrates the position of the genus to be with the 
Antiarcha and not with the Arthrodira as has been suspected by Mr A. S. Woodward." 

Dr Smith Woodward,! however, in a review of the first of Cope's two papers, 
reaffirmed that " the so-called dorsal shield of Holonema is really the ventral shield 
turned the wrong way forwards ; and the genus belongs to the Arthrodira, not to 
the Ostracodermi." Further, that " the limb referred by Professor Cope to Holonema 
is the distal segment of the arm of Bothriolepis, originally named Stenacanthus by 
Leidy." § That he was right in the second statement goes without saying ; and as 

* Proc. Am. States National Museum, vol. xiv. (1891), pp. 447-463. 

t Proc. Am. Phil. Soc., vol. xxx. (1892), pp. 221-229, pis. vii. and viii. 

t Geol. Mag. (3), vol. ix., 1892, pp. 233-235. 

§ It will also be remembered that Agassiz himself was deceived by the corresponding element in Bothriolejns 
major, from the north of Scotland, which he figured, under the name of Placothorax paradoxus, as " un type nouveau 
de la Eamille des Cephalaspides." (1' Foss. vieux gres rouge, p. 134, tab. 30 a, figs. 22, 23.) 


regards the first, there can be no doubt that his explanation of the facts known about 
Holonema is by far the most probable one. 

Consequently we place Holonema in the order Arthrodira, though in what special 
group or family it is still impossible to say. Whether Phyllolepis, the remains of 
which likewise consist of thin plates with a rugose sculpture, also is referable to the 
same order, remains for future discussion. 

Distribution. — Plates certainly referable to Holonema have hitherto been found in 
the United States, and there only in strata of Upper Devonian age. The "Chemung" 
rocks of the States of New York and Pennsylvania are cited as the source of these 
specimens, for though Claypole first stated regarding Professor H. S. Williams's 
specimen that it was from the " Catskill," he also said that " it is right to add that the 
exact horizon of the fossils is a little uncertain. They came, however, from either the 
highest beds of the Chemung or the lowest of the Catskill." 

Dr Smith Woodward, however, in his "Catalogue," part ii. p. 315, mentions that 
"so far as can be determined from the description and imperfect figures, the dermal 
plates from the Devonian of the Eifel named Coccosteus obtusus, Koenen, exhibit much 
resemblance to those of Holonema." I have carefully read Professor von Koenen's 
description # and examined his figures (the latter are certainly imperfect enough), but am 
very doubtful as to the relationship to Holonema of the pieces depicted, especially as 
their radiating striae are said to be " on the under side." The words are : " Diese Platten 
zeigen einerseits eine unregelmassige feine Streifung parallel ihrem Aussenrande, und 
andererseits auf der Unterseite eine feine von einem Puncte ausstrahlende Streifung, 
welche indessen an je einer langen Seite sich steil gegen diese umbiegt." It would 
be necessary to re-examine the originals to come to a definite conclusion on the 

Holonema ornatum, sp. nov., Traquair. PI. II. 

Specific Character. — The marginal radiating ridges are ornamented each with a row 
of minute ganoid tubercles. 

Description. — Only two fragments have occurred. The first of these is represented 
in PI. II. fig. 1, and is evidently a portion of what must have been originally a plate of 
considerable dimensions. It shows the marginal band of ridges very well, but the 
central part of the plate has come off with the counterpart. The band is about 
1 inch in breadth, and consists of ridges which are mostly parallel with each other and 
perpendicular to the free margin of the plate, except at one extremity (the upper in the 
figure), where they are rather oblique. At the opposite extremity of the fragment a 
small part of the central area is seen, and there the ridges become oblique and irregular. 
Where they are regular, from nine to ten may be counted in the ^ inch transversely. 
Each of these ridges is ornamented with one row of minute, rounded, brilliantly ganoid 

* Verh. naturh. Verein. preuss. Rheinl, vol. xliii. (1886), pp. 55-57, with two figures. 


In fig. 2 we have the counterpart of the above-described fragment, which shows a 
slightly larger extent of surface. In the central area of the plate the bone has adhered 
closely to the matrix, so that the impression of the sculpture of that part is shown only 
along a narrow portion on the right side ; and here, as noted above, the ridges become 
oblique and irregular as compared with their tolerably close parallelism in the marginal 

In fig. 3 we have another fragment of quite an irregular shape, namely roughly 
triangular, and measuring 6 inches in greatest length by 3^ in greatest breadth. 
Of the three margins, one (the lower in the figure) is nearly straight and may 
well be considered to have been an original margin, as also may the shorter irregular 
one on the left side ; while the remaining border marks, of course, the line where the 
plate is cut off by the edge of the stone. 

Like the previously described plate, its surface is proportionally thin, its thickness at 
the thickest part being only ^ of an inch. Only towards the short margin is the ex- 
ternal sculpture shown, and this consists mostly of comparatively coarse, plain, reticulat- 
ing ridges, which become finer towards the margin of the plate, and there exhibit 
evident traces of a beaded ornament similar to that which has been noted in the first 
described specimen. 

As the above described two pieces are mere fragments, it is scarcely possible to locate 
them accurately on the body of the fish. I should, however, be inclined to suppose that 
the first (figs. 1 and 2) belonged to the ventral carapace, and is in fact a fragment of the 
anterior or posterior part of the anterior ventro-lateral plate marked b x in text-figure 3. 

Observations. — The sculpture of the first described specimen, which shows promi- 
nently a marginal band of ridges nearly parallel with each other and nearly at right 
angles with the edge of the plate, together with the great proportional thinness of the 
bone, indicate Holonema of Newberry as the genus to which it ought to be referred, 
while the beaded ornament of the ridges supplies us with a tangible specific mark. As 
regards the second example, we note first, that although the marginal band of parallel 
rugge is not visible, owing to the surface of the bone along nearly the whole of the long 
margin having come away with the counterpart, yet the distinct occurrence of the 
minute bead-like tubercles on some of the finer reticulating ridges which are exposed 
indicates that this second plate belongs to the same species as the first. 

Locality. — Bressay, near Lerwick, Shetland ; collected by Dr J. S. Flett. 

Undetermined Fragments. 

The residue of the collection consists of fragmentary dermal plates of fishes, which 
are really unnameable, as their contour affords no clue to their original shape, nor is 
there anything in their sculpture of a really determining character. Of these I have 
figured three examples. 

Fig. 5, PI. I., is an angular fragment showing two overlapped facets and a large 


portion of free surface, which latter is marked with ill-defined rugae, reminding us 
somewhat of the ornament of a scale of Holoptychius. I have, however, no remem- 
brance of having ever seen a Holoptychius scale with overlapped facets so sharply 
marked off from the sculptured area, and ordinarily the overlapped portion of such a 
scale forms an area of much larger proportional extent. 

Fig. 6 is a fragment concerning which still less can be said. It is an absolutely 
irregular five-sided piece seen from the internal or concave aspect, but a certain amount 
of a rugose ornament shows through from the outer surface. These rugse are arranged 
in three groups or areoe, and are, within the limits of each area, parallel to each other 
but at angles to those of adjacent arese, as is well seen in the figure. I cannot venture 
any theory as to the identification of this fragmentary plate. 

The contour of the fragment shown in fig. 7 reminds us superficially of one of the 
posterior ventrolaterals of Coccosteus, but there is no real evidence that this shape 
represents that of the plate in its original entirety. At the upper left-hand corner, 
however, where the surface is better preserved than over the rest of the specimen, 
a few bright and smooth stellate tubercles are seen, which are not unlike those of a 
Coccostean plate. 


[In these Plates all the figures are of the natural size.] 

Plate I. 

Fig. 1. Impression of outer surface of left posterior dorso -lateral plate of an undetermined species of 
Asterolepis, the lower margin being deficient. For orientation see text-figure 1, p. 322. 

Fig. 2. Anterior part of a broken anterior dorso-lateral plate of Asterolepis from which most of the bony 
matter has flaked off. For orientation see text-figure 2, p. 323. 

Fig. 3. Distal segment of pectoral limb of Asterolepis. 

Fig. 4. Articular element of proximal segment of pectoral limb of Asterolepis. 

Figs. 5, 6, 7. Undetermined fragments. 

Plate II. 

Fig. 1. Fragment of a large plate of Holonema omatum, Traq., showing the sculpture along the free 

Fig. 2. Counterpart of the same specimen. 

Fig. 3. Fragment of another large plate, with sculpture more of a reticulated pattern. 


ans. Roy. Soc. Edin r . 

Dr Traquair on Fossil Fishes from Shetland. Plate I. 

Vol. XLVI. 

Fig. i. 

Fig. 6. 

Fig. 2. 


Fig. 4. 

F, 'g- 3- 

Fig. 5. 


£J- /• 

>ans. Roy. Soc. Edln 1 . 

Dr Traquair on Fossil Fishes from Shetland. Plate II. 

Vol. XLVI. 

j^bh /v,o^ 

■^It H>3^ 


XV. — On the Lamellibranch and G-asteropod Fauna found in the Millstone Grit of 
Scotland. By Wheelton Hind, M.D., B.S., F.E.C.S., F.G.S. Communicated by 
Dr J. Horne, F.R.S. (With Two Plates.) 

(MS. received February 12, 1908. Read March 16, 1908. Issued separately September 2, 1908.) 



During the revision of the Midlothian coalfield by the Geological Survey, special 
attention was directed to the plant remains occurring in the Millstone Grit, with the 
view of obtaining evidence bearing on the probable boundary line between the Upper 
and Lower Carboniferous rocks. Mr Tait collected a number of plants from certain 
horizons in this division of the Carboniferous system, which were submitted to Mr 
Kidston, F.R.S. , for determination. Those from the lower part of the Millstone Grit 
were regarded by Mr Kidston as belonging to the Lower Carboniferous flora, and those 
from the upper portion, to the Upper Carboniferous flora. While carrying on his work 
in the field, Mr Tait found certain fossiliferous bands in the basal portion of the Mill- 
stone Grit, charged with lamellibranchs, brachiopods, and other marine organisms, and 
associated with Lower Carboniferous species of plants. These observations were of 
service in tracing the base line of the Upper Carboniferous rocks round a large part of 
the Midlothian basin. It ought to be clearly understood, however, that this boundary 
line is based solely on the evidence supplied by the plants and fishes as determined by 
Mr Kidston and Dr Traquair respectively. 

When the revision of the Midlothian area was completed, certain sections of the 
Millstone Grit in the counties of Linlithgow, Lanark, and Stirling were examined by 
Mr Tait, to see if they might furnish confirmatory evidence of the boundary line 
adopted in Midlothian. Owing to the economic importance of the fireclays in this 
division in Lanarkshire, which has led to extensive boring operations, the sequence of 
the lower part of the strata in that district is now well known. In the basal portion 
there are fossiliferous bands containing marine organisms just as in Midlothian. 
Though Lower Carboniferous plants have not been found in this division in the districts 
of Glenboig, Castlecary, or Torwood Glen, yet they have been met with in beds a little 
above the base of the Millstone Grit at Glasgow, and Upper Carboniferous plants have 
been obtained in the upper part of this division at Castlecary. It seems reasonable to 
infer, therefore, that the lower part of the Millstone Grit in the counties of Lanark, 
Linlithgow, and Stirling is on the same stratigraphical horizon as the basal portion in 
Midlothian with its marine bands and Lower Carboniferous flora. 



While collecting the plants Mr Tait obtained a large number of specimens of 
Lamellibranchiata, which were sent to me for identification. These were found in beds 
exposed in the G-reenfoot Quarry, near Glenboig Station, Lanarkshire (Sheet 31, one- 
inch map) ; in the river Avon, Linlithgowshire (Sheet 31) ; in Torwood Glen, Stirling- 
shire (Sheet 31)*; and in Bilston Burn, Midlothian. At each of these localities 
the lamellibranch fauna was found not far below the line which has been drawn 
between the Upper and Lower Carboniferous floras, in accordance with Dr Kidston's 

I was much surprised to meet with a lamellibranch fauna of which quite 50 per 
cent, of the species were, as far as I could discover, new to Europe, but which resembled 
very closely the lamellibranch fauna of the Coal Measures of Nebraska and Illinois of 
North America. 

The most striking member of the fauna was the shell Prothyris elegans, Meek, this 
being the first occurrence of the genus in the Carboniferous rocks of Great Britain. 
Drevermann has described one species, P. bergica, from the lowest Carboniferous 
rocks of Germany, Zeitsch. der Deutsche Geol. Gesel., vol. xiv., 1902, p. 498, pi. xiv. 
In many cases it was impossible to distinguish any characters sufficient to separate the 
Scotch and American species from each other, but it may be said that the faunas are 
generically the same. Moreover, the lamellibranch fauna shows a marked approach to 
Permian types. 

In North America the fauna is accompanied by Fusulina and Meekella, two fossil 
genera which are absent in the West of Europe ; but both occur in the Spirifer 
mosquensis zone of Russia, a zone which succeeds that of Productus giganteus in that 
country. There does not, however, seem to be the least connection between the fauna 
of the Spirifer mosquensis zone and that under examination. It is important to 
ascertain the horizon at which this interesting fauna occurs in Scotland, that it may 
be known where to look for its representatives elsewhere. At the outset there is a 
great difficulty, and one that must be grappled with in the near future. 

The Lower Limestone series of Scotland contains the same coral fauna as the 
uppermost division of the Carboniferous Limestone of England, or its equivalent farther 
north, the Yoredale series, that is to say, the Dibunophyllum zone. The question at 
once arises, What are the homotaxial equivalents of the Edge Coal series and the Upper 
Limestone series of Scotland, south of the Border ? This question can only be satis- 
factorily settled on palseontological lines. 

There is the most striking difference between the lamellibranch fauna of the Scotch 
beds assigned to the Millstone Grit, and that which characterises the Millstone Grits 
and the Pendleside series of England and Ireland. Here the upper beds of Carbon- 
iferous Limestone, the Oyathaxonia beds, which themselves are the highest subdivisions 
of the Dibunophyllum zone, are succeeded by black shales and limestones with 

* A preliminary statement referring to this collection of lamellibranchs was published in the Summary of Progress 
of the Geological Survey for 1905, pp. 147 and 148. 


Pterinopecten papyraceus, Sow., sp., P. carbonarius, Hind, Posidonomya Becheri, 
Bronn, P. membranacea, M'Coy, Posidoniella Iwvis, Brown, sp., Leiopteria longirostris, 
Hind ; and the black shales which intervened between the different grits of the Mill- 
stone Grit series yield Pterinopecten papyraceus, Sow., sp., Posidoniella Isevis, Brown, 
sp., P. minor, Brown, sp., Sedgwickia attenuata, M'Coy, Sanguinolites ovalis, Hind, 
Mycdina verneuillii, M'Coy, Myalina Flemingi, M'Coy, Schizodus antiquus, Hind, 
Nucula sequalis, Sow., Nuculana stilla, M'Coy, Ctenodonta Isevirostris, Portlock, sp., 
the majority of which have not yet been found in Scotland. On the other hand, I have 
seen goniatites typical of the Upper Carboniferous beds of the Midlands, in the collec- 
tion of Mr Neilson ; Dimorplioceras Looneyi, Phill., sp., D. Gilbertsoni, Phill., sp., 
from shale over the Hosie Limestone ; Glyphioceras reticulatum, Phill., sp., from Gair ; 
G. Phillipsii, Foord and Crick, from Thornliebank ; and from East Kilbride, G. vesica, 
Phill., sp., and Pterinopecten papyraceus, Sow., sp. These species never are found in 
England until the close of the Carboniferous Limestone period, and always occur in beds 
which succeed rocks which have the coral fauna of the Lower Limestone series. I 
mention these facts that the importance of the whole question may be recognised, and 
with the idea that they may possibly lead to a solution of the long open question of the 
difference of the Carboniferous sequence in Scotland and England. 

The brachiopod fauna is not rich in species, but indicates a late period of Carbon- 
iferous time. The special forms of the various species are identical either with those 
that occur in the very fossiliferous quarry 300 feet below the Millstone Grit at 
Congleton Edge, Cheshire, in connection with the Glyphioceras spirale, Phill., sp., beds, 
or in black shales which lie on the upper beds of Carboniferous Limestone in the Upper 
Nidd valley. The most interesting specimens in this connection are the numerous speci- 
mens of Schizophoria resupinata, Mart., sp., where some of the fine radiating lines 
which form the ornament of the shell terminate in small raised points as if there was an 
attempt to form a spine. This character is well marked in the Scotch and Congleton 
Edge shells. There is a species of Lingula which I suspect to be new to science. 

The Cephalopoda are represented by a single species, which I doubtfully refer 
to Glyphioceras Phillipsii, Foord and Crick. The specimen is fragmentary, and is 
a cast of the outer surface of the shell, and does not show any sutures. The 
absence of Cephalopoda at this horizon is very marked compared with their abundance 
throughout the Pendleside series and the shales of the Millstone Grits of England 
and Ireland. 

The Gasteropoda show a strong relation to the North American fauna. Several 
species I regard as identical with those figured and described from the Coal Measures 
of Nebraska. 

Mr Tait writes me that he " obtained no plants in the marine beds in the Greenfoot 
Quarry, in the Avon section, nor in Torwood Glen, but in Garngad Eoad, Glasgow, 
plants belonging to the Lower Flora occur in beds 50 to 100 feet above the Castlecary 
Limestone." " From the upper part of the Millstone Grit at Castlecary, plants belong- 



ing to the Upper Flora were obtained, so that the horizon at which the flora changes 
is almost the same level as that in Midlothian. I think that all the lamellibranchs 

Boulder Clay. 

J Dark Shales. 

Thin Coal. 

Shales and Fireclay, nodular. 
Fireclay, with Sandstone ribs 
and irony nodules (" Bullet" of Glenboig). 

Fireclay (mined). 

Interval concealed. 


Dark grey and blue stained Shales. 

Grey Shales with rib or nodules of Clayband Ironstone. ^ 

Cementstone, varies much in thickness, sometimes in two bands / 

Shales with hard rib of Kingle or Gannister. 

Sandy Shales and Fireclay. 

:.■ Faky Sandstone 

Soft gritty white and yellowish 
false-bedded Sandstone in thick beds, 
wrought for moulding sand. 

(Anthracomya truncata, sp. nov. 

Dark Shales J Pterinopecten papyraceus. 

Impure ferruginous Limestone (Elonichthys pectinatus. 

(This is regarded as the position 

of the Castlecary Limestone.) 

f Zaphrentis Omaliiisi, M. Ed. & H. 
Densiphyllum charlestonense , Thorn. 
Limatuiina alternate/,. 
Myalina Verneuillii. 
Posidoniella Icevis. 
Prothyris elegans. 
Protoschizodus curtus. 
Solenom orpha cylin drica. 
Solenomya brevis. 
Tellinomorpha Hindii. 
Ptychomphalus Marcouianus. 

Fig. 1. — Vertical Section of Strata at Greenfoot Quarry, near Gain Farm, 3 miles N. of Coatbridge. 

were obtained from beds containing a Lower Carboniferous flora, and just below the 
point where the Upper and Lower floras meet." 

Certain species which were found to occur in the Basement beds of the Bristol 



Clay Shale, lower part with marine shells. \ I 
Shelly Calcareous IRONSTONE. y--< 

Coal 2"-4". 

fCtenodonta Icevirostris. 

Edmondia excentrica. 
, , sulcata. 

Limatuiina alternata. 

Nuculana attenuata. 

Posidoniella eorrugata. 

Protoscldzodus curtus. 

Pterinopecten Whitei. 

Sanguinolites occidentalis. 

Schizodus W'heeleri. 

Tellinomorpha Hindii. 
filadodus striatus (tooth). 

Fig. 2.— Section on the North Bank of the River Avon, West of the Railway 
Viaduct, 1\ miles W.N.W. of Linlithgow. Scale, 1 inch = 30 feet. 

Coalfield were sent to me by Mr Bolton for my opinion. Amongst them I recognised 
three species which were common to those beds and the Millstone Grit of Scotland, 
namely, Palaeolima retifera, Shumard, Grammatodon tenuistriatus, M. and H., sp., 



and a new form named by Mr Bolton Tellinamorpha Hindii. His paper is published 
in the Quart. Journ. Geol. Soc, vol. lxiii. p. 445. 

The annexed vertical sections illustrate the order of succession of the strata at 
the localities where the lamellibranchs were obtained ; the one at Greenfoot Quarry- 
has been kindly supplied by Mr Hinxman, the second and third on the river Avon 







Gap with Sandstone. 

Coal 10". 


Ironstone with Productus. 

Gap. Probably 
includes a good 
deal of Sandstone. 

Impure ferruginous Limestone or \ 

Shale, with Ironstone nodules. Marinefossils. / 

Curdly IKONSTONE, calcareous in parts, impure. 
\Clay shale with marine fossils. 

Ctenodonta latvirostris. 
Edmondia excentrica. 

„ sulcata. 
Limatulina alternata. 
Nuculana attenuate/,. 

„ Icevistriata. 
Grammatodon tenuistriata. 
Posidoniella Icevis. 
Protoschizodus curtus. 
Pterinopecten papyraceus. 

,, Whitei 

Sanguinolites occidentalis. 
Schizodus Wheeleri. 
Tellinomorpha Hindii. 

/Bituminous Shale with fish scales and plant remains 

(Horizon of Levenseat Oilshale). 
Castlecary Limestone, formerly quarried and mined. 

Fig. 3. — Section on the Banks of the River Avon, East of the Railway Viaduct a mile N. of 
Manuel Junction. One-inch Map 31. Scale, 1 inch = 30 feet. 

by Mr Clough, the fourth at Tor wood Glen by Mr Tait, and the fifth in Bilston 
Burn by Mr Clough, and I owe a debt of thanks to Dr Lee for references to the 
discovery of Prothyris by Drevermann in Germany. 

It is certain that the fossiliferous horizon indicated in the shorter section lies 
above those shown in the larger one, but, owing to the variability of the beds 
and an obscure fault, its exact distance above these horizons is doubtful. It cannot, 



however, be lower than the thick fireclay, and it may be 30 or 40 feet above the 
top of it. 







Fireclay. ( Productus. 

„,„ n .. „ ,- . BeUerophon. 
^Shelly Cement— Brachiopoda? rolled.. J Derhya 

\Tliin Coal. | Ling'ula. 

\Fireclay. I Schizodns Wheeleri. 


Mkmx kk 




I Lint 
| Deri 


/Shale, with fossils. ' Prc( i uc 't us . 
Barren Cement. BeUerophon. 

K KKKKKKK1 \ Tnin Coal - \Lamellibranchs. 

v " Fireclay. 

Sandstone ends off and the two fireclays at one place come together. 




Sandstone, clayey. 



l \] ' : ' :.:]'■}':■ -"^ 

ii i.i i_T~i~r 

= \ Fireclay 





Shelly Cement— Brachiopoda? rolled. Fossils all Brachiopoda. 



/Poor Sandy Cement. 

r- Bluish Shale. 

( Ctenodonta Icevirostris. 

I Shale Nucula ceguaiis. 

" JCastlecaky Limestone. \ Suculana attenuata. 

I Derby a. 

Ctenodonta Icevirostris. 
Edmondia excentrica. 
, , Lyelli. 
,, sulcata. 
Limatulina alternata. 
Modiola subelliptica. 
( Myalina redesdalensis. 
,, Verneuillii. 
Nuculana attenuata. 
Protoschizodus curtus. 

„ suvcequalis. occidentalis. 
Schizodus Wheeleri. 
Tellinomorpha Hindii. 
Entails Meekianum. 
\Ptychomphalus Marcouianus. 

Fig. 4. — Section of Millstone Grit in Torwood Glen, 2J miles N. W. of Larbert, Stirlingshire. 
One-inch Map 31. Scale, 1 inch = 30 feet. 

Specimens T2143 B — 2150 B , from Glencryan, 1^- miles S.S.E. of Cumbernauld, One 
inch Map 31, include the species — 

Posidoniella l&vis, or P. corrugata. 

Specimens T2176 B — 2178 B , from Glencryan — as above, but lower in the series — 
contain 2 Edmondia Lyelli. 

Specimens T2214 B -2243 B are from the Fireclay Works, f mile a little N. of E. of 
Castlecary Railway Station, Stirlingshire. One-inch Map 31. 

The following species were obtained here: Sanguinolites, sp., Palseolima retifera, 
Shumarcl, sp., ? Edmondia Lyelli, Hind. This mine is sunk to the Castlecary Lime- 



stone, so there is no doubt the beds from which these fossils were obtained overlie it, 
though the bed has not been recognised in situ. 

Upper Carboniferous plants were got in this mine at 270 feet above the Castlecary 
Limestone. The species determined by Mr Kidston are Sigillaria elegans, Sternb. 
(abundant), Lepidodendron obovatum, Sternb., Lepidodendron aculeatum, Sternb. 


Fireclay 3". \ 
Slaty C< >AL 3". 
Pale grey Clay Shale, with scattered 
hard nodules weathering yellow, Calc 1 / 
becoming more abundant at base. f ' 
Hard Rib, with hard clayey matrix -v / 
and many hard, yellow, weathering \/ 
nodules, slightly calcareous. J 


= Blaes. 

■ ~ /Red Slaty Ironstone 

Clay Shale, mottled pale grey and red. 
Mottled red, white, and yellow Sandstone. 

]\1 assive 
in parts. 


Fireclay, rather pale grey. 

/Bastard Ironstone. 
/ Blaes. 

Ironstone Rib with shells. 
\ Blaes. 

Calcareous Blaes. 

Dark grey slialy LIMESTONE. 

Dark grey Clay Shale. 


Blaes 3". 



Sandstone and some Fakes. 

** ' : -T"^ ! l/Ccarss pebbly Rib with Clay galls Pyritss 
- i .,<• remains 

~ : — [and plant 

j I tULj \ yS c [ castlecary Limestone (top post). 

r Chonetes, sp. 

Derbya, sp. 

IAngula mytiloides 
(cf. M artinia glabra). 

Orbiculoidea nitida. 

Productus (cf. longispinus). 

Pugnax pugnus. 

1 Seminula ambigua. 

? Spiriferina cristata. 

Aviculopecten obliqnus. 
,, regidaris. 

? ,, negtectus. 

Clenodonta Iwvirostris. 

Edmondia senilis. 

? Grammatodon tenuistriata. 

Myalina Venieuillii. 

Nuculana attenuata. 

Pterinnpecten j>apyrace\is. 

Sanguinolites occidentalis. 

Euphemus d'Orbignyi. 

Goniatite (cf. Glyphioceras diadema). 
^ Orthoceras, sp. 

Fig. 5. — Vertical Section of Strata in Bilston Burn, near Pol ton, Midlothian. 


Specific Descriptions. 

PalsBolima retifera, Shumard, sp., 1858. (PI. I. fig. 1.) 

Lima retifera, Shumard, 1858, Trans. St Louis Acad. Set., vol. i. p. 214. 

,, Geinitz ? , 1866, Garb, und Dyas in Nebraska, p. 36, pi. ii. figs. 20, 21. 
,, „ Meek and Hayden, 1872, Fin. Rep. U.S. Geol. Surv. Nebraska, p. 188, pi. ix. 

fig. 5. 
Palxolima retifera, Bolton, 1907, Q. J. Geol. Soc, p. 462, pi. xxx. fig. 10. 

Specific Characters. — Shell small, oblique, triangularly ovate, very moderately 
gibbose. The anterior border only slightly curved, the inferior broadly rounded, the 
posterior oblique and almost straight. The hinge line is short and straight. The 
umbones are subcentral and sma]l. The anterior ear is small, compressed, triangular ; 
the posterior very similar, and almost the same size. The valve is flattened near 
the anterior border and expands gradually so that the greatest convexity is near the 
posterior border, where the valve becomes bent on itself so as to form a flattened 
posterior surface. 

Interior.— The shell is ornamented with many angular radiating ribs, somewhat 


irregular in distribution, and those towards the posterior margin stouter than those in 
front. Occasionally ribs may bifurcate, but more frequently new ones arise between 
any pair. These ribs are crossed by fine concentric lines of growth. The ears are 
almost smooth. 

Dimensions.— PI. I. fig. 1. Specimen No. T2218 B , a right valve, measures: 
antero-posteriorly, 12 mm. ; dorso-ventrally, 14 mm. 

Localities. — Bed in pit shaft at Castlecary Fireclay Works. 

Observations. — There is one specimen of this shell in the collection which I have 
referred to Shumard's species. The very elaborate description and excellent figures 
given by Meek and Hayden leave nothing to be said in addition. The British species 
most closely related is Paleeolima simplex, Phillips. This shell is less triangular, 
comparatively broader, and has fewer but broader and less angular radiating ribs. 
Meek and Hayden, following Shumard, refer their shell to Lima, but I have shown 
that the Carboniferous forms differ in certain details [Brit. Carb. Lamell., vol. ii. p. 39) 
from typical Limas, and created the genus to receive them. The occurrence of the 
genus in rocks of Millstone Grit age in Scotland somewhat extends our knowledge of 
the vertical range of the genus. 

Limatulina alternata, M'Coy, sp., 1844. (PL I. figs. 9, 10.) 

Limatulina alternata, Hind, 1903, Brit. Carb. Lamell. (Pal. Soc), vol. ii. p. 39, pi. xix. 
figs. 7-10, 12. 

Seven examples of this shell have been obtained, five of which occurred in an 
impure ferruginous limestone or cement in the river Avon, between the Railway 
Viaduct and the outcrop of the Castlecary Limestone, in the county of Linlithgow. 
Some of the specimens are full grown and show no tendency to dwarfing or senile 
characters. For further details, my work on Carboniferous lamellibranchs (op. supra 
cit.) may be consulted. The figured specimens are numbered Tl888 B and Tl818 B . 

Pterinopecten Whitei, Meek, sp., 1872. (PI. I. figs. 11-13.) 

Aviculopecten Whitei, Meek, 1872, Fin. Rep. U.S. Geol. Surv. Nebraska, p. 195, pi. iv. 
fig. 11, a, b, c. 

Specific Characters. — Shell below medium size, subordinately quadrate, the left 
valve moderately convex, the right valve flattened. The anterior and lower borders 
rounded. The posterior bluntly rounded below, slightly sinuous above. The hinge 
line straight, meeting the posterior border with a well-marked angle. The umbones 
are small, pointed, flattened, not elevated, placed far forward. The anterior ear is 
small, somewhat angular, well defined from the valve by a deep sulcus, deeper in the 
right valve than in the left. There is no posterior ear, but the upper and posterior 
part is compressed and subalate. The surface is ornamented by flattened, flexuous, 
radiating ribs, irregular in size, being crowded and narrow anteriorly, broader and 


more widely separated over the middle portion of the valve, becoming very fine and 
close on the compressed part of the valve, so as to appear obsolete to the naked eye. 
These radiating ribs are crossed by concentric, irregularly distributed lines and rugae 
of growth. 

Dimensions. — PL I. fig. 11. Specimen No. T2982 B measures: antero-posteriorly, 
15 mm. ; dorso-ventrally, 14 mm. 

Localities. — Linlithgow, river Avon, E. side of the Kailway Viaduct, 35 to 40 
feet above the Castlecary Limestone. 

Observations. — A number of specimens of this species have been obtained, but from 
only one of the localities. The species agrees with the description given by Meek 
(op. supra cit.), but is somewhat larger than his figured specimens. In the adult state 
the posterior margin becomes sinuous above, a fact not indicated in the small American 
specimen. The species has some affinity to P. carbonarius, Hind, but the ribs are 
broader and flatter and more flexuous, and more interrupted by concentric lines of 
growth. The distribution of P. carbonarius extends from fairly low down in the 
Pendleside series to the middle of the Coal Measures. 

Pterinopecten papyraceus, Sow., sp. 

I refer two fragments to this well-known species, one from the Linlithgow and the 
other from the Gair locality. Undoubted specimens, however, were obtained from the 
Bilston Burn. I feel fairly satisfied with my diagnosis, even in the fragmentary state 
of the specimens. It is at this horizon that I should expect the species to occur in 
force. In the Midlands it occurs immediately above the upper beds of Carboniferous 
Limestone or the beds denoted by the coral genera Cyathaxonia and Amplexi 

Aviculopecten carboniferus, Stevens, sp., 1858. (PL I. fig. 14.) 

Pecten carboniferus, Stevens, 1858, Amer. Journ. Sci. and Arts, vol. xxx. p 261. 

„ Broadheadii, Swallow, 1862, Trans. St Louis Acad. Set., vol. ii. p. 97. 

,, Hawni, Geinitz, 1866, Carb. und Dyas in Nebraska, p. 36, pi. ii. fig. 19, a, b. 
Aviculopecten carboniferus, Meek and Hayden, Fin. Rep. U.S. Geol. Surv. Nebraska, p. 193, pi. iv. 

fig. 8 ; pi. ix. fig. 4, a, b. 

Specific Characters. — Body of the shell triangularly ovate, convex, greatest length 
and breadth equal. The hinge line is moderately elongate, the umbones pointed, raised, 
and central. The inferior border is broadly rounded, the margin interrupted by the 
projection of the ribs as pointed processes. The anterior ear is large, compressed, 
somewhat rolled, its upper border projecting and its anterior edge prolonged into a 
point ; the posterior ear larger than the anterior, pointed, its margin falcate. The 
ears are separated from the body of the valve by a deep concave sulcus. 

Exterior. — The surface is ornamented with few (15) distant, regular, raised, rounded 
ribs, occasionally nodulose when they are crossed by transverse lines of growth. These 
ribs project beyond the lower margin in the form of spines. 



Dimensions. — PI. I. fig. 14. Specimen No. 15,731, a left valve, measures: antero- 
posteriorly, 18 mm. ; dorsoventrally, 17 mm. 

Locality. — Millstone Grit of Hazel Hill, 5 miles N.E. of Pately Bridge, Yorkshire ; 
and possibly Congleton Edge Quarry, Cheshire. 

A single specimen, a left valve, is in the collection of the Geological Survey, Jermyn 
Street, from the Millstone Grit of Hazel Hill, Yorkshire. 

Observations. — The specimen which I now refer to this species is fairly well preserved, 
but has the tips of each ear wanting. At the lower margin the ribs are seen to project 
as spines, but the matrix is too hard and coarse to make out whether the spines curved 
upwards as they are shown to do in Meek and Hayden's figure. The only other British 
Carboniferous example which is known to have had projecting ribs is A. Murchisoni 
(Brit. Carb. Lamell., vol. ii. pi. xiv. fig. 4), a species which more closely resembles the 
American shell than any other of the genus. Mkek and Hayden describe the right 
valve as nearly flat, with surface markings as in the left valve but more obscure. 

Two fragments in my possession from the Congleton Edge Quarry, upper part of 
the Pendleside series, probably belong to the species, although not present in the Scotch 
collection. I have inserted the description of A. carboniferus here because it is a 
member of the American fauna. 

Aviculopecten obliquus, sp. nov. (PI. I. fig. 8.) 

Specific Characters. — Shell small, oblique ; body of the valve ovate, acute. The 
right valve convex. The anterior margin is short and convex. The inferior descends 
and is broadly convex. The postero-inferior part of the valve produced and bluntly 
rounded. The line which forms the margin from the umbo to the postero-inferior angle 
is almost straight, and defines the valve from the posterior ear. The hinge line is 
straight, long, produced posteriorly into an acute point. The umbones are small, 
pointed, level with the hinge line, and placed very slightly anterior to the centre of the 
hinge line. The anterior ear is depressed, rolled, comparatively large and deep, separated 
from the valve by a fold and notch. The posterior ear is much compressed, triangular, 
pointed, its margin falcate. 
Interior. — Not exposed. 

Exterior. — The right valve is ornamented by many thick, nodulose, radiating ribs, 
between each pair of which linear plain ribs are intercalated. The anterior ear has a 
few radiating ribs ; the posterior, fine concentric lines of growth. 

Dimensions. — PL I. fig. 8. Specimen No. T4507 B , a right valve, measures : antero- 
posterior^, 12 mm. ; dorso-ventrally, 12 mm. 

Locality. — Midlothian, Bilston Burn near Polton, bed 27 to 30 feet above Castlecary 

Observations. — Unfortunately only right valves of this species have been found. 
The nodular ornament is very well marked. One specimen, which consists of the fossil 
and its counterpart, shows an almost perfect shell ; the other is a cast of the exterior. 


The species has somewhat the characters of the ornament of A. Murchisoni; the 
obliquity of the valve and the deep posterior ear are not characteristic of that species. 
In A. Murchisoni, the right valve is flattened, and the ribs are much less nodular than 
in the left valve. 

Aviculopecten regularis, sp. nov. (PL 1. figs. 2, 3.) 

Specific Characters. — Shell of moderate size, the left valve gibbose, the right 
flattened. The body of the valve triangular, the lower margin broadly rounded. The 
hinge line straight, equal in length to the greatest transverse diameter. Umbones 
small, pointed, subcentral. The ears are depressed, especially the anterior ; the posterior 
has its upper margin produced to a point, and its posterior margin falcate. 

Exterior. — The right valve has its anterior ear radially ribbed and its posterior 
almost smooth, but almost obsolete distant radiating lines can be seen with a microscope. 
The rest of the valve has fine flattened, close radiating ribs, becoming thicker as they 
approach the margin. The left valve has coarse, raised linear ribs ; between each pair, 
fine secondary ribs appear towards the inferior margin. Concentric lines of growth are 
visible in the umbonal region. The ears of the left valve are not well seen. 

Dimensions. — PI. I. fig. 3. Specimen No. T4514 B measures: antero-posteriorly, 
16 mm. ; dorso-ventrally, 13 mm. 

Locality. — Midlothian, Bilston Burn near Polton, bed 27 to 30 feet above the 
Castlecary Limestone. 

Observations. — Two specimens, a right and a left valve, which I think belong to 
each other, have been obtained, neither of which is, unfortunately, perfect. The marking 
of the left valve is well preserved. I have been unable to refer the specimens to any 
described species, and I have been obliged to consider it as a new species. 

? Aviculopecten neglectus, Geinitz, sp., 1866. (PI. I. figs. 4-7.) 

Pecten neglectus, Geinitz, 1866, Garb, tmd Dyas in Nebraska, p. 33, T. ii. fig. 17. 
Aviculopecten neglectus, Meek, 1867, Amer. Journ. Sci. and Arts, vol. xliv., sec. ser., p. 183. 

„ „ Meek, 1872, Fin. Rep. U.S. Geol. Surv. Nebraska, p. 193, pi. ix. fig. 1, a, b. 

„ ,, Meek and Worthen, 1873, Geol. Surv. Illinois, vol. v. p. 589, pi. 261, fig. 7. 

Specific Characters. — Shell small, its body subovate, acute, only moderately gibbose. 
The margin is regularly rounded below. The hinge line straight, shorter than the 
greatest transverse diameter. The umbo is small, pointed, not elevated, subcentral. 
The anterior ear in the right valve is small and flattened, triangular, separated from the 
body of the valve by a groove and notch. It is ornamented by strong radiating ribs. 
The posterior ear is triangular, depressed, with fine radiating markings, its margin 

Exterior. — The right valve has its body almost smooth, with fine concentric 
markings. The left valve is almost smooth in its juvenile condition, but becomes 


finely radially striate near its lower margin in the more recent part of the valves. 
Shell very thin. 

Dimensions. — Specimen No. T4503 B . a right valve, measures : antero-posteriorly, 
10 mm. ; dorso-ventrally, 10 mm. PI. I. fig. 5. 

Locality. — Midlothian, Bilston Burn near Pol ton, bed 27 to 30 feet above the 
Castlecary Limestone. 

Observations. — Several specimens of the right valve of this species, but only two of 
the left, have been obtained. Comparing the specimens with the figures given by Meek 
in the report on Nebraska (op. supra cit), it will be seen that he gives a very perfect 
figure of the right valve, with which our specimens agree ; but it would seem that he 
copied Geinitz's figure of the left valve, which I have reasons for thinking may have 
been a right valve with the ears imperfectly exposed, because the left-hand ear in 
that specimen has a falcate margin and should therefore be the posterior. Amongst 
the specimens from the Bilston Burn is a left valve (fig. 4, PI. I.) which shows in its 
umbonal region a surface free from radiating ribs, but towards the margin ribs gradually 
arise and become close and numerous. The specimen is a left valve, and I have sus- 
pected that it is the left valve of the smooth right valves which agree with the 
American right valve in every essential character. 

In vol. v. of the Geological Survey report on Illinois, Worthen gives the figure 
of the hinge plate, showing a number of parallel cartilage pits, closer on the anterior 
than on the posterior side, and a central cartilage cavity. If he is correct, the shell 
cannot be referred to Aviculoyecten, but will require a new genus to be created for it. 

Meek figures a left valve with radial ribs, from Bed C, Nebraska, in which 
? A. neglectus occurs, as A. coxanus, M. and W. It is possible that this may be the 
left valve, and that I am dealing with partially decorticated specimens, and am wrong 
in thinking that the valve was not radially ribbed in the umbonal region. 

Posidoniella Isevis, Brown, sp., 1841. (PI. I. fig. 15.) 

For synonymy, vide Hind, Pal. Soc, 1897, Brit. Garb. Lamell., p. 94. 

I have referred certain small shells to this species, which is extremely common in 
the shales of the Millstone Grit in the Midlands of England. The specimens are not 
good, and are small and crushed, but on the whole fairly distinctive of the species; it 
seems to have been very rare in Scotch localities, and dwarfed. 

Myalina Verneuillii, M'Coy, sp., 1854. (PI. I. fig. 16.) 

For synonymy, vide Hind, Pal. Soc, 1897, Brit. Garb. Lamell., p. 115. 

I have referred some fragments of a large shell with a broad alate posterior end to 
this species, but they are also very suggestive of M. redesdalensis. Some smaller 
specimens one can positively refer to M'Coy's species, which has been obtained in beds 
of the Millstone Grit series near Marsden, on Pule Hill, Yorkshire. M. Fleminyi occurs 


with this species at Pule Hill, and I have recognised that species in a collection from 
Garngad Road made by Mr J. Neilson of Glasgow. 

Locality. — 9 to 10 feet above the Castlecary Limestone, Torwood, Stirlingshire; 
Greenfoot Quarry, near Gain Farm, Dumbartonshire ; and Bilston Burn, Midlothian. 

Nucula gibbotsa, Flem., vide Hind, Pal. Soc, 1897, Brit. Garb. LamelL, p. 178. PI. I. 

figs. 18, 19. 
Nuculana attenuata, Flem., sp., vide Hind, Pal. Soc, 1897, Brit. Garb. Lamell., p. 195. 
Ctenodonta Ixvirostris, Portl., sp., vide Hind, Pal. Soc, Brit. Garb. Lamell., pp. 183, 210; pt. ii. 

p. 164. 

These three species occur together at many horizons in the Carboniferous sequence, 
from the Calciferous Sandstone series to the Coal Measures. 

It is interesting to note that each genus is represented in the Coal Measures of 
North America by a closely allied species — Nuculana (Yoldia) subsulcata, Nuculana 
bellistriata, Ctenodonta (allorisma) reflexa. 

Nuculana Isevistriata, Meek and Worthen, sp. (PI. I. fig. 17.) 
For synonymy, vide Hind, Pal. Soc, 1897, Brit. Garb. Lamell., p. 205. 

This species is found in the St Louis group of the Lower Carboniferous beds of 
North America, but in Scotland it is very common at Waulkmill Glen in the Upper 
Limestone series. It seems to be rare in the Millstone Grit localities. 

Locality. — River Avon, between the outcrop of the Castlecary Limestone and the 
Railway Viaduct. 

Modiola subelliptica, Meek, 1867. (PI. I. fig. 20.) 

Clidophorus (Pleurophorus) occidentalis, Geinitz, 1866, Garb, und Dyas in Nebraska, p. 23, pi. ii. 

fig. 6. 
Pleurophorus subellipticus, Meek, 1867, Amer. Journ. Sei. and Arts., vol. xliv., new ser., p. 181. 
Modiola 1 subelliptica, Meek, Fin. Rep. U.S. Geol. Sura. Nebraska, 1872, p. 211, pi. x. fig. 5. 

Specific Characters. — Shell small, transversely triangular, comparatively convex. 
The anterior end is small, narrow and rounded. The posterior end comprises about 
four-fifths of the valve. The anterior border is narrow and elliptical. The lower border 
descends slowly, meeting the posterior with a broad, blunt curve. The posterior margin 
is very oblique and somewhat convex in contour, making an obtuse angle above with 
the hinge line. The latter straight, equalling in length almost half the antero-posterior 
diameter. The umbones are small, placed far forward, but not terminal. Well marked, 
rounded, oblique swelling, which gradually becomes obsolete, extends from the umbo to 
the postero-inferior angle. Above the ridge the dorsal slope is rapidly compressed, and 
below it there is a well-marked byssal compression. 

Interior. — Not exposed. 

Exterior. — The surface is adorned by close, fine, concentric lines of growth. 

Dimensions. — Specimen No. Tl712 B measures: antero-posteriorly, 7 mm.; dorso- 
ventrally, 3 mm. 


Locality. — Torwood Glen, Stirlingshire, 8 to 10 feet above the Castlecary 

Observations. — A single specimen of this shell has occurred, a left valve, and I have 
referred it to Meek's species. The American shell reaches about 25 mm. in its trans- 
verse diameter, and therefore we must regard the Scotch specimen as immature. 

Grammatodon tenuistriata, Meek and Worthen, sp., 1872. (PI. I. fig. 21.) 

Macrodon tenuistriata, Meek and Worthen, 1867, Proc. Chicago Acad. Set., i. p. 17. 

Area striata, Geinitz, 1866, Garb, und Dyas in Nebraska, p. 20, pi. i. fig. 32. 

Macrodon tenuistriata, Meek and Hayden, 1872, Fin. Rep. U.S. Geol. Surv. Nebraska, p. 207, 

pi. x. fig. 20, a, b. 
Parallelodon tenuistriata, Bolton. 1907, Q. J. Geol. Soc, p. 460, pi. xxx. fig. 12. 

Specific Characters. — Shell small ; somewhat transverely oblong, gibbose. The 
anterior end is bluntly rounded, the antero-superior angle well marked. The inferior 
and superior borders sub-parallel. The hinge line straight, the lower margin very 
slightly convex, deeply indented above its centre for the byssus. The posterior end 
obliquely truncate, almost straight. The umbones are turned, raised, incurved, and 
pointed, placed at the junction of the anterior and middle thirds of the hinge line. The 
body of the valve is convex, deeply indented towards the margin of the byssal sulcus. 
The dorsal slope is as much compressed as to be concave. It is separated from the rest 
of the valve by a well-marked angular ridge, which extends from the umbo to the 
postero-inferior angle. 

Interior. — Normal. 

Exterior. — The general surface of the valve is ornamented by the decussation of the 
concentric lines of growth by close radiating lines. The latter become so much stronger 
on the dorsal slope as to form strong radiating ridges. 

Dimensions. — PI. I. fig. 21. Specimen No. Tl812 B measures : antero-posteriorly, 
12 mm. ; dorso-ventrally, 7 mm. 

Locality. — Linlithgow, right bank of river Avon between the Railway Viaduct and 
the outcrop of the Castlecary Limestone, and in the Bilston Burn section, Midlothian. 

Observations. — G. tenuistriata, M. and H., is closely allied to G. semicostatus, 
M'Coy, of which it is probably a mutation. It is, however, much less transverse, deeper 
and more gibbose. 

Only two examples of the species are in the collection. G. semicostatus is very 
abundant at a horizon a very little below that in which G. tenuistriata occurs, namely, 
below the Linn Spout Limestone. Mr H. Woods has shown {Ann. Mag. Nat. Hist., 
1899, ser. 7, vol. iii. p. 47) that the name Grammatodon, Meek and Hayden, is prior 
to Parallelodon of the same authors, and must replace it. It appears that Macrodon of 
Lycett, founded for a Jurassic shell, was occupied for a genus of fishes and cannot 
therefore be used, and that there is no essential difference in the type of the hinge of 
the Carboniferous and .Jurassic forms. 


Protoschizodus curtus, Meek and Worthen, sp., 1866. (PI. I. figs. 22, 23.) 

Schizodus curtus, Meek, 1866, Proc. Chicago Acad. Sci., vol. i. p. 18. 

,, rossicus, Geinitz, 1866, C'arb. und Dyas in Nebraska, p. 18, Tab. i. fig. 28. 
,, curtus, Meek and Worthen, 1872, Fin. Rep. U.S. Geol. Sure. Nebraska, p. 208, pi. x. 
fig. 13, a, b, c, dl, e. 

Specific Characters. — Shell triangularly suborbicular, moderately gibbose, almost 
equilateral. The anterior margin almost straight above, oblique, bluntly rounded below 
in the inferior border, which is broadly convex. The posterior margin is narrowed, 
straight, obliquely truncate. Hinge line arched short. The umbones are well marked, 
raised, pointed, moderately gibbose, and central. Passing from the umbo obliquely to 
the postero-inferior angle is a well-marked angular ridge which separates the convex 
portion of the valve from the rapidly compressed dorsal slope. 

Interior. — Muscle scars normal. 

Exterior. — The surface is almost smooth, but obscure concentric lines of growth are 
apparent near the lower margin of the valve. 

Dimensions. — PI. I. fig. 22. Specimen No. Tl982 B , a right valve, measures: 
antero-posteriorly, 16 mm. ; dorso-ventrally, 15 mm. 

Localities.— Stirlingshire, Torwood Glen, 2^ miles N.W. of Larbert, 9 to 10 feet 
above Castlecary Limestone ; Linlithgow, river Avon, right bank, between Railway 
Viaduct and outcrop of Castlecary Limestone ; Greenfoot Moulding Sand Quarry, near 
Gain Farm, Dumbartonshire. 

Observations. — Protoschizodus curtus resembles very closely P. sub&qualis, de 
Kon, from the upper beds of Carboniferous Limestone of Thorpe Cloud, Upper Dibuno- 
phyllum zone ; but the Scotch examples are less gibbose, more triangular, and have a 
more oblique, straighter anterior border. The umbones are more subcentral and more 
elevated. Meek and Worthen recognised the close resemblance between the American 
shell and Schizodus rotundatus of the Permian. I think they were right in separating 
the species for several reasons. In the first place, S. rotundatus was established on a 
single example, and I believe has not been figured since King's monograph was published 
in 1841 ; and we learn from the note on the explanation of the plate that "the figure 
makes the specimen more perfect than it really is." It is important, however, to note 
the close relationship between the three forms. 

I have referred the species to Protoschizodus on account of its general shape and 
contour. This species is fairly common in at least two of the Scotch localities. 

Schizodus Wheeleri, Swallow, sp., 1862. (PI. I. figs. 22-28.) 

Cypricardia ? Wheeleri, Swallow, 1862, Trans. St Louis Acad., vol. i. p. 96. 

Schizodus obscurus, Geinitz, 1866, Carb. und Dyas in Nebraska, p. 20, Tab. i. figs. 30, 31. 

„ Wheeleri, Meek and Hayrlen, Fin. Rep. U.S. Geol. Surv. Nebraska, p. 209, pi. x. fig. a, 
b,c,d, (e,f1). 

Specific Characters. — Shell moderately gibbose ; dorsal slope much compressed, 


transversely subovate. The anterior margin is deep and regularly rounded, passing 
with uninterrupted curve into the lower border, which is also convex, but the segment 
of a larger curve than the anterior margin. The posterior end is narrowed, obliquely 
truncate, almost straight. The postero-inferior angle acute, the postero-superior obtuse. 
The hinge line is short, very slightly arcuate. The umbones are large, gibbose, elevated 
above the margin, incurved, and subcentral. Passing obliquely from the umbo to the 
postero-inferior angle is a well-marked subangular ridge which separates the dorsal slope 
from the convex portion of the valve. 

Interior. — The hinge plate is normal. 

Exterior. — The surface is ornamented in front by regular lines of growth, which 
become less marked over the posterior portion of the valve. 

Dimensions. — PL I. fig. 24. Specimen No. Tl865 B , a right valve, measures: 
antero-posteriorly, 26 mm. ; dorso-ventrally, 20 mm. 

Locality. — Stirlingshire, Torwood Glen, 2^ miles N.W. of Larbert, in a bed 9 to 10 
feet above Castlecary Limestone; Linlithgow, river Avon, about 100 yards below the 
Mill, right bank, at the bend where stream takes a northerly direction, about \ mile N. 
of Avonbank. 

Observations. — This species is represented by a larger number of individuals than 
any other in the collection. It is seen that amongst them there is a fair amount of 
variation. The same thing is observed in the figures of the American types figured by 
Meek. We can even match the larger form (fig. 1,/), doubtfully referred to the species 
by Meek, amongst the Scotch specimens. There is certainly a much closer resemblance 
to the Permian Schizodus obscurus than to any other British species ; but it would 
seem to connect S. axiniformis, which ranges from the upper part of the Carboniferous 
Limestone series (Redesdale ironstone) to the Coal Measures (Coalbrookdale), with the 
later form. From the study of a series it would appear that, with age, the angularity 
of the oblique ridge tends to disappear, and the posterior end becomes produced. 

? Anthracomya truncata, sp. nov. (PI. I. figs. 29, 30.) 

Specific Characters. — Shell small, triangular, compressed, rapidly expanded 
posteriorly, greatest vertical and horizontal diameter equal. The anterior end is 
short, its margin rounded. The inferior border descends rapidly, joining the long, 
straight, obliquely truncate posterior border by an elliptical curve. The hinge line is 
straight ; the longest transverse diameter of the valve forms a well-marked obtuse angle 
with the posterior border. The umbones are small, slightly elevated and swollen, placed 
at the junction of the anterior and middle thirds of the hinge line. The valve is 
obliquely swollen, forming a well-defined but gradually diminishing rounded ridge 
passing from the umbo towards the postero-inferior angle. 
Interior. — Not exposed. 


Exterior. — The surface is almost smooth, but under the lens close, fine concentric 
lines of growth, with here and there a stronger line, the latter being more conspicuous 
in the umbonal region. The periostracum apparently thick. 

Dimensions. — Specimen No. T2743 B measures : antero-posteriorly, 5 mm. ; dorso- 
ventrally, 5 mm. 

Locality. — In black micaceous shale in drain in bottom of Greenfoot Quarry, near 
Gain Farm, 3 miles N. of Coatbridge. (Position of Castlecary Limestone.) 

Observations. — Two small pieces of black shale labelled T2742 B and T2743 B are 
covered with black shiny shells and fragments which are new to me. The shells are 
probably somewhat crushed, but they have the general shape and contour of Anthra- 
comya, to which genus I am led to refer them. They differ from A. valenciensis in 
being more triangular and less globular, and in having a well-marked oblique rounded 
ridge and a very rapidly expanded posterior extremity. The comparatively thick 
periostracum is a fact in favour of affinity of this small shell to Anthracomya. One 
slab shows four fairly well-preserved specimens of left valves, the other a practically 
perfect specimen of the right valve. 

For its size this species is more rapidly expanded posteriorly than any of its genus, 
and foreshadows the character of the large Anthracomya Adamsii and its variety 
A. Adamsii, var. expansa, of the middle portion of the Coal Measures of England. A 
slab labelled T2156 B , from a dark shale on top of a 2-inch coal resting on the Glenboig 
Fireclay Seam, Fireclay Mine, Glenboig. 

Edmondia excentrica, sp. nov. (PI. II. figs. 31, 32.) 

Specific Characters. — Shell small, only moderately gibbose, subrotundate, quadrate, 
moderately transverse, inequilateral. The anterior end is bluntly rounded, the inferior 
margin broadly convex, the posterior blunt, convex, somewhat broader than the anterior. 
The hinge line arcuate. The umbones are small, incurved, placed in front of the 
middle portion of the valve. 

Exterior. — The surface is ornamented with concentric lines and rugae of growth, 
and are arranged somewhat obliquely to the long axis of the valve. 

Dimensions. — PI. II. fig. 32. Specimen No. Tl945 B measures: antero-posteriorly, 
18 mm. ; dorso-ventrally, 12 mm. 

Localities. — Stirlingshire, Torwood Glen, 2|- miles N.W. of Larbert, 9 feet above 
Castlecary Limestone ; river Avon, \ mile S.E. of Inveravon, Linlithgowshire ; Glen- 
cryan, If miles S.S.E. of Cumbernauld, Dumbartonshire. 

Obse7'vatio?is. — A typical Edmondia, with an affinity to E. senilis, but of quite a 
different habit of growth, hence its creation into a species. 

The Edmondia reflexa of Meek, from the Coal Measures of North America, is a 
more transverse shell with more regular and less rugose markings. 



Edmondia Lyellii, Hind, 1899. (PL II. fig. 35.) 
For synonymy, vide Hind, Pal. Soc, 1899, Brit. Garb. Lamell., vol. i. p. 300. 

A fragment of what I take to be the right valve of this species was obtained at 
Torwood Glen, Stirlingshire. 

Edmondia sulcata, Phillips, sp. (PI. II. fig. 33.) 
For synonymy, vide Hind, Pal. Soc, 1899, Brit. Garb. Lamell., p. 318. 

A single specimen of this shell has been obtained from the Millstone Grit horizon 
at river Avon, between the Railway Viaduct and the Castlecary Limestone, Linlithgow. 
The specimen, though crushed, is well preserved, and not to be mistaken. It is a 
right valve, and shows in places the radiating lines of tubercles with which the posterior 
portion of the valve was covered. Another better specimen was got on the same 
horizon at Torwood Glen, Stirlingshire. Other examples were got at Gain Quarry, 
Dumbartonshire. The figured example, No. Tl858 B , is from the river Avon, on west 
side of the Railway Viaduct, 100 yards below the Mill. 

Sanguinolites occidentalis, Meek and Hayden, sp., 1858. (PL II. figs. 36-38.) 

Pleurophorus occidentalis, Meek and Hayden, 1858, Trans. Albany Instit., iv. 

„ ,, Meek and Hayden, 1864, Pal. Upper Missouri, p. 35, pi. i. fig. 11a. 

Clidophorus Pallasi, Geinitz pars., 1866, Garb, und Dyas in Nebraska, p. 23, pi. ii. fig. 3. 
Pleurophorus occidentalis, Meek and Hayden, 1872, Fin. Rep. U.S. Geol. Surv. Nebraska, p. 212, 
pi. x. fig. 12. 

Specific Characters. — Shell ovate, oblong, broader behind than in front, obliquely 
gibbose. The anterior end is short, its margin rounded. The upper and lower borders 
subparallel. The posterior margin obliquely subtruncate, junction of lower and posterior 
borders bluntly rounded. The postero-superior angle well marked, obtuse. The 
umbones are small, pointed, directed forwards, only very slightly raised above the 
hinge line and placed far forwards. Passing obliquely from the umbo to the postero- 
inferior angle is a bluntly rounded ridge which separates a broad and compressed dorsal 
slope from the narrow and somewhat compressed body of the valve. The edge of the 
upper margin is thickened posteriorly, and forms a ridge ; between it and the oblique 
ridge are two well-marked oblique lines which extend to the posterior border. 
Escutcheon well marked. 

Interior. — Normal. Hinge plate apparently edentulous. 

Exterior. — The surface is ornamented by somewhat irregular rugose but fine lines 
of growth, which become much less obvious on the dorsal slope. 

Dimensions.— -PI. II. fig. 37. Specimen No. T2216 B , a left valve, measures: 
antero-posteriorly, 20 mm. ; dorso-ventrally, 9 mm. 

Localities. — Linlithgow, river Avon, right bank, between Railway Viaduct and the 
outcrop of the Castlecary Limestone; Stirlingshire, Torwood Glen, 2|- miles N.W. of 


Larbert, 9 to 10 feet above the Castlecary Limestone. Specimen No. T2216 B is from 
the pit shaft of Steens Fireclay Works, f mile E. of Castlecary Railway Station ; 
Bilston Burn section, Midlothian. 

Observations. — I have referred this species to the genus Sanguinolites rather than 
Pleurophorus, as there are no signs of the characteristic cardinal teeth of the latter 
genus. I think there can be little doubt of the relation of the shell to the American 
species. I have compared the suite of specimens with Sanguinolites ovalis, Hind, but 
the two radiating ridges on the dorsal slope at once separate it from this species. The 
figure of the specimens from Nebraska (op. supra cit.) is from the cast of an interior, 
copied from Geinitz's work. It is shown to have three curious curved lines passing 
from the front of the umbo to the lower border. I cannot think that these are natural, 
and I have never seen any similar structure in a sanguinolitiform shell. The species 
is represented in the collection by seven examples, amongst which is a fairly well- 
preserved testiferous example, and a cast of the left valve (PI. II. fig. 38). 

This species belongs to the group of Sanguinolites tricostatus, but it is considerably 
less transverse than that species. Externally there is a strong resemblance to the 
Pleurophorus elegans of the Permian. 

? Allorisma refiexa, Meek. (PI. II. fig. 39.) 
Allorisma (Sedgwickia) refiexa, Meek, Fin. Rep. U.S. Geol. Sum. Nebraska, p. 217, pi. iv. fig. 15. 

A small slab, No. T2177 B , is covered with a very strongly marked shell, of which, 
however, no single specimen is perfect. I refer it provisionally to Allorisma refiexa, 
Meek, with the description of which it seems to agree. 

I have doubts whether Meek's generic diagnosis is correct, as the shell has not the 
general shape usual in Allorisma. 

The dimensions of one of the Scotch shells, a right valve, are : antero-posteriorly, 
20 mm. ; dorso-ventrally, 8 mm. 

Locality. — Glencryan, If mile S.S.E. of Cumbernauld, Dumbartonshire, 10 to 20 
yards down stream from the fault, and on the under side of the waterfall. 

Tellinomorpha Hindu, Bolton, 1907. (PI. II. figs. 40-42.) 
Tellinomorpha Hindii, Bolton, 1907, Q. Journ. Geol. Soc, vol. lxiii. p. 460, pi. xxx. fig 8, a, b. 

Specific Characters. — Shell somewhat transverse, gibbose, with a much-compressed 
dorsal slope, inequilateral. The anterior end is short and narrowed, its margin rounded. 
The inferior border is gently curved, and meeting the posterior at a well-marked angle. 
The posterior border is obliquely truncate, almost straight, moderately short, making 
an obtuse angle with the hinge line, the latter straight. The umbones are gibbose, 
elevated, incurved, and situated in front of the middle of the valve. Lunule large and 
broad. Escutcheon narrow, but well marked. From the umbo to the postero-inferior 


angle of the valve extends a blunt subangular ridge, which divides the convex portion 
of the valve from the concavity of the dorsal slope. 

Interior. — Not known. 

Exterior. — The surface of the valve is ornamented with numerous concentric, 
irregular rugae and lines of growth, which tend to become obsolete on the dorsal slope. 

Dimensions. — Specimen No. T2450 B measures: antero-posteriorly, 11 mm.; dorso- 
ventrally, 7 mm. ; gibbosity of valve, 3 mm. 

Localities. — Dumbartonshire, Greenfoot Moulding Sand Quarry, near Gain Farm, 
3 miles N. of Coatbridge ; Stirlingshire, Torwood Glen, 1\ miles N.W. of Larbert, 9 to 
10 feet above the Castlecary Limestone; Linlithgow, river Avon, 35 to 40 feet above 
the Castlecary Limestone. 

Observations. — This species is represented in the collection by a very large number 
of individuals, by far the largest number of which came from the Gain Quarry. The 
specimens occur in ferruginous nodules which lie above the Upper Cementstone. The 
shell has a distinctive appearance, being very rugose for so small a shell. The narrowed 
truncate posterior end and well-compressed dorsal slope, comparatively large lunule and 
escutcheon, indicate fairly conclusively, even in the absence of details of the hinge and 
interior, the generic affinity of the species. The genus Tellinomya is represented in 
the Coal Measures of North America by Tellinomya [Allorisma, Sedgivickia) granosa 
of Shumard, and probably the Allorisma [Sedgivickia) subelegans of Meek. The 
former species is very much larger than the one just described, but the latter agrees 
with it very well as to size. From the figure and descriptions, the ornament appears 
to be regular, and the shell is described by Meek as rather compressed, a character 
which does not suit the Scotch species. With a strong lens, I think the remains of 
the characteristic tubercles with which the surface of shells belonging to the genus are 
covered are to be seen. Since writing the description I recognised the species in a 
collection made by Mr H. Bolton from the base of the Coal Measures of Bristol, who 
has published a description and figures {pp. supra cit.). He appears doubtful of the 
genus, and suggests Parallelodon, but I cannot accept this view. The small amount 
of the hinge plate that I have seen has none of the characters of that genus. The 
antero-superior angle of the shell is quite unlike the arciform type, and is rounded, and 
not square and compressed. 

Solenomorpha cylindrica, sp. nov. (PL II. figs. 43, 44.) 

Specific Characters. — Shell small, narrow, transverse, very inequilateral, dorsal and 
ventral margins subparallel. The anterior end is comparatively elongate, narrow ; its 
anterior superior angle almost pointed, the border elliptically curved. The posterior 
end is truncate, blunt, not much narrowed, making a rounded angle below with the 
inferior margin, and a slightly obtuse angle with the hinge line above. The umbones 
are compressed, elongate, and not elevated. The dorsal slope is slightly compressed, 


bounded below by an obscure ridge which continues from the umbo to the postero- 
inferior angle. 

Interior. — Not observed. 

Exterior. — -The surface appears to be almost smooth. 

Dimensions. — Specimen No. 2441, a left valve, measures: antero-posteriorly, 
9 mm. ; dorso-ventrally, 3 mm. (PL II. fig. 44.) 

Locality. — Dumbartonshire, Greenfoot Moulding Sand Quarry, near Gain Farm, 
3 miles N. of Coatbridge. 

Observations. — I have referred these specimens to the genus Solenomorpha on 
account of the elongate form and the peculiar elongated anterior end and the narrow 
compressed umbones. The posterior end is broader than is usual in species of this 
genus. The American Coal Measures species (Solenopsis) Solenomorpha solenoides of 
Geinitz has the normal condition of a narrowed posterior end and a somewhat less 
prolonged anterior end. 

Solenomya brevis, sp. nov. (PL II. figs. 45-47.) 

Specific Characters. — Shell small, transverse, tumid, lenticular, narrowed at both 
extremities, very inequilateral. The anterior end long, its margin elliptical. The 
inferior margin is convex ; the posterior truncate, rounded. The hinge line is gently 
arched. The umbones are inconspicuous, elongate and adpressed, placed far back so 
that the posterior end is short and rapidly compressed, especially above. 

Interior. — Normal. 

Exterior. — Almost smooth, with very fine concentric striae of growth. 

Dimensions. — PL II. fig. 46. Specimen No. T2718 B measures: antero-posteriorly, 
16 mm. ; dorso-ventrally, 8 mm. ; gibbosity of valve, 3 mm. 

Locality. — Dumbartonshire, Greenfoot Moulding Sand Quarry, near Gain Farm, 
3 miles N. of Coatbridge. 

Observations. — This species belongs to the group of Solenomya, which does not 
possess radiating ribs passing across the middle part of the valves. The absence of 
these ribs and the peculiar boat-shaped valve at once distinguish the species from 
S. cylindrica, which occurs at the same horizon. Its nearest congener is S. excisa, 
but this is a much lamer and more transverse shell. 


Solenomya cylindrica, sp. nov. (PL II. figs. 60, 61.) 

Specific Characters. — Shell small, narrow, transversely elongate, gibbose, very 
inequilateral. Anterior end long, its margin bluntly rounded. The hinge line and 
inferior border are almost parallel. The posterior end narrower than the rest of the 
valve, its margin elliptical. The umbones are small, not elevated, excavated posteriorly, 
and placed in the posterior fourth of the hinge line. 

Interior. — Normal. 


Exterior. — The surface is almost smooth, but with the microscope fine concentric 
lines of growth are visible. The valve is crossed by several radiating obscure flattened 
ribs, apparent even in casts. 

Dimensions. — PI. II. fig. 60. Specimen No. Tl984 B , a right valve, measures : antero- 
posteriorly, 19 mm; dorso-ventrally, 7 mm. 

Locality. — Dumbartonshire, Greenfoot Moulding Sand Quarry, near Gain Farm, 
3 miles N. of Coatbridge. 

Observations. — The species differs from S. costellata, M'Coy, being much more 
gibbose and comparatively more transverse and much narrower. The radiating ribs 
are much less obvious. Three specimens are present in the collection. It is possible 
that the species is a mutation of S. costellata, for there is no doubt that the two forms 
are closely related, but the differences seem to me to be sufficiently strong to be 
indicated by a new term. Apparently the genus is represented in the Coal Measures 
of Illinois, for the name of the genus occurs twice in the list on page 126 of the Final 
Report of U.S. Geological Survey of Nebraska, but no mention is made of any species 
in the text. 

Genus, Prothyris, Meek, 1869. 

Prothyris, Meek, 1869, Proc. Acad. Nat. Sci. Philad., p. 172. 

„ Meek, 1871, Amer. Jourri. Conch., vol. vii. p. 5, pi. i. fig. 3. 

„ Meek, 1872, Fin. Rep. U.S. Geol. Surv. Nebraska, p. 223, pi. x. fig. 9, a, b. 

„ Hall, 1885, Geol. Surv. N. York, Pal., vol. v. pt. i., Lamellibr. ii., p. xl. 

„ Miller, 1889, N. Amer. Geol. and Pal., p. 504. 

„ Whidborne, Pal. Soc, 1890, Devonian Fauna, p. 86. 

Generic Characters. — Shell compressed, narrow, transversely elongate, with a small 
ear-like process from the antero-posterior angle, separated from the shell by an angular 
ridge, probably for the byssus. Dorsal slope compressed. 

Observations. — The genus is easily recognised by the peculiar-shaped process at its 
anterior end. The genus was founded by Meek for a shell from the Coal Measures of 
Nebraska. The original account (op. supra cit.) was very meagre, and unaccompanied 
by any figures. However, in 1871 he gave figures, and, I presume, a more elaborate 
description, and again in 1872 (op. supra cit.). In North America the genus was sub- 
sequently found to occur as low as the Hamilton series, where it is represented by two 
species. Two others are known from the Chemnung and one from the Waverley Sand- 
stone. In England, the Rev. G. F. Whidborne has described three species from 
the Marwood and Pilton beds of North Devon. 

The internal characters are not known. I think it probable that the ear-shaped 
process contained the anterior abductor muscle, and the hollow separating it from the 
rest of the shell was for a byssus. Fischer doubtfully places the genus in the Solemdae. 
Apart from the anterior ear-like process, the shell has a strong resemblance to some 
species of Sanguinolites. 


Prothyris elegans, Meek, 1871. (PL II. figs. 48-50.) 

Prothyris elegans, Meek, 1871, Ajner. Journ. Conch., vol. vii. p. 5, pi. i. fig. 3. 

,, ,, Meek and Hayden, 1872, Fin. Rep. U.S. Geol. Surv. Nebraska, p. 223, 

pi. x. fig. 9, a, b. 

Specific Characters. — Shell moderately compressed, transversely oblong, dorsal and 
ventral margins straight and parallel. The anterior superior angle is expanded into a 
small, pear-shaped process, not extending below the upper third of the border, the 
broad end pointing downwards and forwards. This process is separated from the rest 
of the shell by a raised ridge, which is continued below as the anterior margin, which is 
bluntly rounded. The posterior margin is obliquely subtruncate and almost straight. 
The postero- inferior angle bluntly rounded ; the postero-superior angle almost obtuse. 
The umbones very small, depressed, and placed very far forwards : they would be 
terminal were it not for the ear-like expansion. The upper margin of each valve is 
compressed by a shallow groove immediately below the hinge line. Another groove 
passes obliquely from the region of the umbo towards the posterior margin some little 
distance below. Below these two grooves is a rounded ridge, which becomes broader 
as it approaches the posterior end. Below this the valve is very gently convex. 

Interior. — Unknown. 

Exterior. — That portion of the valve near the upper margin and the dorsal slope is 
almost smooth. Nearer the ventral border there are fine close lines of growth, parallel 
with the margins. 

Dimensions. — PI. II. fig. 48. Specimen No. T2505 B , a right valve, measures : 
antero-posteriorly, 17 mm. ; dorso-ventrally, 5 mm. 

Locality. — Dumbartonshire, Greenfoot Moulding Sand Quarry, near Gain Farm, 
3 miles N. of Coatbridge. Horizon : Millstone Grit of Scotland. 

Observations. — Although the Rev. G. F. Whidborne has recorded three species of 
the genus Prothyris from the Devonian rocks of England, it has not been previously 
found in the Carboniferous beds of Great Britain [vide ante, p. 332). The type of the 
genus is Prothyris elegans, a species which occurs in the Coal Measures of Nebraska 
and Illinois. I have been unable to note any specific difference between the American 
and Scotch specimens after careful comparison with a series from Nebraska in my 
collection. Many examples have been obtained by Mr. Tait from the locality named 
above — fortunately right and left valves. In some the ear-like process which is 
characteristic of the genus has broken away. The left valve is represented by the 
fossil and its counterpart, the latter showing the ear very perfectly. I have seen other 
examples in the possession of Mr. J. Smith of Dairy. 


Loxonema nanum, de Koninck, 1881. (PL II. fig. 51.) 
Loxonema nanum, de Koninck, 1881, Ann. Mils. Roy. d'Hist. Belg., tome vi. p. 50, pi. iv. figs. 45, 46. 
Specific Characters. — Shell very small, elongate ; spire consisting of 8 to 9 whorls. 


Regularly convex. Suture lines well depressed ; often a slight ridge immediately above 
the suture. 

Dimensions. — Specimen No. T2182 B measures: height, 4"5 mm. ; breadth, 1*5 mm. 

Localities. — Moulding Sand Quarry, Garngad Road, Glasgow ; Dumbartonshire, 
Greenfoot Moulding Sand Quarry, near Gain Farm, 3 miles N. of Coatbridge. 

Observations. — Two specimens of this minute shell have been obtained. It is 
almost characterless, and I refer it on this account to the very small shell named by 
de Koninck L. nanum. 

Naticopsis brevispira, de Ryckholt, sp. (PI. II. figs. 52, 53.) 

I described this species as occurring in a marine band below the Gin Mine Coal of 
the North Staffordshire coalfield (Q. J. Geol. Soc, vol. lxi. p. 533, pi. xxxv. fig. 10). 
A single specimen has been obtained from Torwood Glen, Stirlingshire. 

Macrocheilina, sp. (PI. II. fig. 59.) 

Fragments of a small elongate shell with a spire of five whorls have been obtained, 
but it is impossible to determine the species. The height is 4 '5 mm. The specimen is 
numbered Tl929 B . 

Locality. — Linlithgowshire ; river Avon, W. of the Railway Viaduct, about 
100 yards below the Mill and \ mile N. of Avonbank. 

Ptychomphalus Marcouianus, Geinitz, sp., 1866. (PL II. fig. 55.) 

Pleurotomaria Marcouiana, Geinitz, 1866, Carb. und Dyas in Nebraska, 1866, p. 10, Tab. i. fig. 10. 
„ „ Meek and Hayden, 1872, Fin. Rep. U.S. Geol. Surv. Nebraska, p. 223, 

pi. xi. fig. 8. 

Specific Characters. — Shell small, cone rapidly expanding, spire of 5 to 6 whorls, of 
which the larger comprises nearly the whole of the shell. The upper portion of the 
whorl is convex, and in the last whorl projects beyond the two keels of the band of the 
sinus, so that the latter is finally not marginal. It is ornamented with regular close- 
set spiral ridges. The portion of the whorl above the band of the sinus (i.e. nearer the 
apex) is flattened, and in the early whorls obliterates the sinus. The ornament consists 
of spiral bands which are moniliform internally, the points becoming small in each 
successive row till they become obsolete. 

Dimensions. — PI. II. fig. 55. Specimen No. Tl779 B measures: height, 6 mm.; 
breadth, 7 mm. 

Localities. — Stirlingshire, Torwood Glen, 2|- miles N.W. of Larbert ; Dumbarton- 
shire, Greenfoot Quarry, near Gain Farm, 3 miles N. of Coatbridge. 

Observations. — I have referred these specimens to Geinitz's species, trusting to the 
figures in his and Meek's works. I have been fortunate enough to obtain for study a very 


beautifully preserved example which shows the character of the ornament very perfectly. 
I can find no figure of any British Carboniferous shell which in any way agrees. 

Entalis Meekianum, Geinitz, sp., 1866. (PL II. fig. 54.) 

Dentalium Meekianum, Geinitz, 1866, Garb, und Dyas in Nebraska, p. 13, Tab. i. fig. 20. 

„ ,, Meek and Hayden, 1872, Fin. Rep. U.S. Geol. Surv. Nebraska, p. 224, pi. xi. 

fig. 16, a, b. 

Specific Characters. — Shell a conical tube with a gentle curvature ; aperture circular. 
Lines of growth fine, with irregular deeper sulci. 

Locality. — Stirlingshire, Torwood Glen, 2| miles N.W. of Larbert. Tl765 B . 

Observations. — A single specimen of this shell has occurred. The exterior is not 
well preserved, but seems to be smooth. It may be decorticated, however. Dentalium 
Meekianum, Geinitz, from the Coal Measures of Nebraska, is said to have fine moderately 
distinct lines of growth passing very obliquely round the shell, with here and there a 
deeper sulcus. Of this character I can see no trace on the Scotch specimen. 

A species of the genus is not uncommon in the marine beds of Congleton Edge, 
300 feet below the third grit. It is much larger than the Scotch specimen, but of 
course this is not of specific value, and it may be said that only the narrow terminal 
portion of the latter is preserved. The Congleton Edge specimen has no longitudinal 
striae, but has almost concentric lines of growth with irregular deeper sulci in places, 
thus corresponding to Meek's description, to which I have alluded above. Meek draws 
attention to the fact that Geinitz represented the lines of growth in his figures as if 
they were spiral — a fact which he states to be incorrect. I have followed de Koninck: 
in using the generic name of Entalis, this genus possessing a slit on the dorsal surface 
of the posterior or narrow end which is absent in Dentalium. 

Euphemus d'Orbigniji, Port., sp. (PL II. figs. 57, 58.) 

Bellerophon d'Orbiynyi, Portlock, 1843, Rep. Geol. Londonderry, p. 401, pi. xxix. fig. 12. 
1 Bellerophon carbonarius, Cox, 1857, Kentucky Geol. Rep., vol. iii p. 562. 

„ ,, Geinitz, 1866, Garb, und Dyas in Nebraska, p. 6, Tab. i. fig. 8. 

„ „ Meek and Hayden, 1872, Fin. Rep. U.S. Geol. Surv. Nebraska, p. 224, 

pi. iv. fig. 16 ; pi. xi. fig. 11, a, b, c. 
Euphemus d'Orbiynyi, de Koninck, 1833, Ann. Mus. Roy. d'Hist. Nat. Belg., tome viii. p. 156, 
pi. xlii. bis, fig. 5; pi. xliii. figs. 9-12; pi. lxii. figs. 10-12. 

Specific Characters. — Shell gibbose, umbilicus small, sides compressed. Aperture, 
transverse subtruncate. The surface is ornamented by about twenty-four spiral ridges 
separated by concave spiral sulci, which become obsolete over the greater portion of the 
terminal whorl, and are closer together on the side of the umbilical slope, 

Localities. — Stirlingshire, Torwood Glen, 1\ miles N.W. of Larbert ; Linlithgow- 
shire, river Avon ; Midlothian, Bilston Burn section. 



Dimensions. — PI. II. fig. 57. Specimen No. Tl756 B measures: height, 16 mm.; 
transversely, 16 mm. 

Observations. — Portlock states that he described his species from a single example, 
which he stated was obtained from Shale in Tyrone. De Koninck referred several 
specimens from the Carboniferous Limestone of Vise" to this species, and gave numerous 
figures. I am able to detect no specific difference between the descriptions and figures 
of Bellero'phon carbonarius, Cox, from the Coal Measures of Nebraska, and the Scotch 
examples, and have come to the conclusion that both should be referred to Poktlock's 

Euphemus, sp. 

Two poor specimens, with a finely reticulate ornament, were obtained at Torwood 
Glen, Stirlingshire. I don't think that they belong to any described species, but the 
shells are much too imperfect to serve as types for a new species. Nos. Tl776 B , Tl799 B . 

Belleroplion Marcouianus, Geinitz, 1866. (PI. II. fig. 56.) 

Bellerophon Marcouianus, Geinitz, 1866, Garb, und D;/as in Nebraska, p. 7, Tab. i. fig. 12. 

,, ,, Meek and Hayden, 1872, Fin. Rep. U.S. Geol. Surv. Nebraska, p. 226, 

pi. xi. fig. 13, a, b. 

Observations. — A single species from Torwood Glen must, I think, be referred to the 
species figured by Conrad from the Lower Coal Measures of Western Virginia. Meek 
figures a fragment only, but this shows the characteristic keel. The Scotch specimen is 
very badly preserved : the strong median imbricated ridge is well seen, but the rest of 
the shell, though not crushed, is so encrusted by mineral matter that the surface of the 
valve is not to be made out. I know no figured shell except the North American species 
which has such a strongly marked and peculiarly ornamented median ridge. m 



Table showing the Lamellibranch Fauna of the Coal Measures of 
Nebraska, and the occurrence of the same Species in Europe. 





Base of Bristol 
Coal Measures. 


Limatnlina alternata, M'Coy, sp., 


Palxolima retifera, Shumard, . 

• 1 x 


Entolium aviculatum, Swallow, sp., 

1 x 

x + + 

Nucula Beyrichi, V. Schaur. 1 . 


„ gibbosa, Flem., 


,, ventricosa, Hall, . 


Yoldia subscitula, M. and H., . 


Nuculana Isevistriata, M. and W., sp., 



,, attenuata, Flem., sp., 


Ctenodonta Ixvirostris, Portlock, sp., 


Grammatodon tenuistriata, M. and W., sp. 





Protoschizodus curtus, M. and W., sp., 



Schizodus Wheeleri, Swallow, sp., 



x t:;: 

Avicula longa, Geinitz, 


„ ? sulcata, Geinitz, 


Pseudomonotis radialis, Phill., sp., 


x I 

Myalina Swallovi, M'Chesney, . 


,, subquadrata, Shumard, 


,, perattenuata, M. and W., 


,, Verneuillii, M'Coy, 


Avictdopecten oecidentalis, Shumard, sp., 


,, neglectus, Geinitz, sp., 



,, carboniferus, Stevens, sp., 





,, coxanus, M. and W., . 


,, obliquus, sp. nov., 


,, regtdaris, sp. nov.. 


Pterinopecten papyraceus, Sow., sp., 



„ Wliitei, Meek, sp., 



Posidoniella Ixvis, Brown, sp., . 




Aviculopinna americana, Meek, 


Pinna peracula, Shumard, 


Modiola ? subelliptica, Meek, . 


PI europlwrus oblongus, Meek, . 


x ! 

Sanguinolites oecidentalis, M. and H 

, sp., 



Edmondia reflexa, Meek, . 



,, glabra, Meek, . 


,, 7iebrasce>/sis, Geinitz, sp., 




., subtruncata, Meek, . 


„ aspinwallensis, Meek, 


,, Lyellii, Hind, . 


,, sulcata, Phill., sp., . 


Allorisma reflexa, Meek, . 



,, Oeinitzi, Meek, 


,, subelegans, Meek, 


,, granosa, Shumard, sp., 


„ subcuneata, M. and H., 



Tellinomorphu Hindii, Bolton, 



Solenomorplia cylindrica, sp. nov., 


Solenomya brevis, sp. nov., 


,, cylindrica, sp. nov., 


Prothyris elegans, Meek, . 



Solenopsis solenoides, Geinitz., sp., 


* Tschernyschew, " Obercarbonischen Brachiopoden d. Ural u. d. Timan," Mem. Com. Ge'ol. Russie, 1902. 

t Stuckenberg, " Die Fauna der obercarbon. Suite des Wolgadurbruches bei Samara," Mem. Com. Ge'ol. Russie, 
1905, Livr. 23. 

t Jakowlew, " Die Fauna der oberen Abtheilung der Palaeozoischen Ablagerungen im Donetz Bassin : Die 
Lamellibranchiaten," Mem. Com. Ge'ol. Russie, 1903. 



Since completing and presenting my account of the lamellibranch fauna in the 
Millstone Grit in Scotland, the following specimens have been collected by the officers 
of the Geological Survey from the cores of the Plean Bore near Stirling. 

Edmondia nebrascencis, Geinitz, sp., 1866. (PI. II. fig. 34.) 

Astarte nebrascencis, Geinitz, 1866, Garb, unci Dyas in Nebraska, p. 16, Tab. i. fig. 25. 
Edmondial ,. Meek, 1872, Fin. Rep. U.S. Geo!. Surv. Nebraska, p. 214, pi. x. fig. 8, a, b. 

Miller, 1889, N. Amer. Geol. and Pal., p. 479. 

Specific Characters. — Shell slightly transverse, subovate, only moderately gibbose, 
unequilateral. The anterior end is short, its margin regularly rounded. The inferior 
margin is broadly curved. The posterior border is bluntly rounded. The hinge line 
is slightly arcuate. The umbones are small, slightly elevated, placed at the junction 
of the anterior and middle thirds of the hino-e li ne< 

Interior. — Not yet examined. 

Exterior. — The surface is ornamented by concentric, fine, raised, fairly equidistant 

. linear ridges which separate moderately wide concentric sulci which are finely linear 

transversely. This marking is more characteristic in the new parts of the valve, 

towards the inferior margin ; in the umbonal or juvenile part of the shell the lines 

and grooves are crowded. Under the microscope very fine radiating lines are seen. 

Dimensions. — A right valve from Nebraska in my collection measures : antero- 
posteriorly, 22 mm. ; dorso-ventrally, 17 mm. 

Locality. — Stirlingshire, Sheet 24 N.W., Eosehill diamond bore, \\ mile E. of 

Observations. — Some half-dozen specimens of this species have been obtained from 
the diamond bore at Eosehill. E. nebrascensis is very closely related to E. M'Coyii, 
Hind, but it has some slight differences in the ornament. The concentric lines are more 
sharply linear and finer, especially in the older part of the shell, and in one specimen 
I made out the radiating markings mentioned by Meek in his species. In general 
shape there is no real difference between the species, and they are very closely allied, 
but E. M^Coyii is the more gibbose of the two. I have a series of fine specimens from 
Nebraska, and have been able to compare these with the Scotch examples and the series 
of shells which served me for the study of my species E. M l Coijii. Unfortunately the 
Scotch examples are very badly preserved ; but specimens T£|£f£?, PI. II. fig. 34, show 
the characteristic marking of the valve. The series from Eosehill show a greater tendency 
to variation of the ornament in the older and younger portions of the shell than my 
American series. Sanguinolites occidentals occurs in the Shale with E. nebrascensis. 



Plate I. 

Fig. 1. Paleeolima retifera, Shumard, sp. x 3/2. Page 337. 

Figs. 2, 3. Aviculopecten regularis, sp. nov. x 3/2. Page 341. 

Figs. 4-6. Aviculopecten neglectus, Geinitz, sp. x 2. Page 341. 

Fig. 7. „ „ „ 4. Page 341. 

Fig. 8. Aviculopecten obliquus, sp. nov. x 2. Page 340. 

Figs. 9, 10. Limatulina alternata, M'Coy, sp. Page 338. 

Figs. 11, 12. Aviculopecten Whitei, Meek, right and left valves. Page 338. 

Fig. 13. „ „ x 3/2. Page 338. 

Fig. 14*. Aviculopecten carboniferus, Stevens, sp. x 3/2. Page 339. 

Fig. 15. Posidoniella leevis, Brown, sp., a left valve, x 3. Page 342. 

Fig. 16. Myalina Verneuillii, M'Coy, sp. Page 342. 

Fig. 17. Nuculana Ixvistriata, Meek and Worthen, sp. x 3/2. Page 343. 

Figs. 18, 19. Nucula gibbosa, Flem. x 4 Page 343. 

Fig. 20. Modiola subelliptica, Meek. x 3. Page 343. 

Fig. 21. Grammatodon tenuistriata, Meek and Worthen, sp. x 21. Page 344. 

Figs. 22, 23. Protoschizodus curtus, Meek and "Worthen, sp. x 3/2. Page 345. 

Figs. 25-28. Schizodus Wheeleri, Swallow, sp. Page 345. 

Figs. 29, 30. 1 Anthracomya truncata, sp. nov. Page 346. 

All specimens are from Scotland, and in the Museum of the Geological Survey of Scotland, with the 
exception of *. This specimen is from Pately Bridge, Yorkshire, and is in the Collection of the Geological 
Survey, Jermyn St. 

Plate II. 

Figs. 31, 32. Edmondia excentrica, sp. nov. Page 347. 

Figs. 33. ,, sulcata, Phillips, sp. Page 348. 

Fig. 34. „ nebrascensis, Geinitz, sp. Page 358. 

Fig. 35. „ Lyelli, Hind. Page 348. 

Figs. 36-38. Sanguinolites occidentalism Meek and Hayden, sp. Page 348. 

Fig. 39. 1 Allorisma reflexa, Meek. Page 349. 

Figs. 40-42. Tellinomorplia Hindii, Bolton, x 3/2. Page 349. 

Figs. 43, 44. Solenomorplia cylindrica, sp. nov. x 3. • Page 350. 

F'ig. 45-47. Solenomya brevis, sp. nov. Figs. 45 and 47 x 2. Page 351. 

Figs. 48-50. Prothyris elegans, Meek. x 3/2. Page 353. 

Fig. 51. Loxonema nanum, de Koninck. x 3. Page 353. 

Figs. 52, 53. Naticopsis brevispira, de Ryckholt, sp. x 3. Page 354. 

Fig. 54. Enialis MeeJcianum, Geinitz, sp. x 3/2. Page 355. 

Fig. 55. Ptychomphalus Marcouianus, Geinitz. sp. x 2. Page 354. 

Fig. 56. Bellerophon Marcouianus, Geinitz. x 3/2. Page 356. 

Figs. 57, 58. Euphemus d'Orbignyi, Portlock, sp. x 3/2. Page 355. 

Fig. 59. Macrocheilina, sp. Page 354. 

Figs. 60, 61. Solenomya cylindrica, sp. nov. Page 351. 

AU the figured specimens are from Scotland, and are in the Museum of the Geological Survey of Scotland. 


Iraxis.Iloy: Soc.Mkf 

1 X % 





2 x 3 /- 

3 x 3 / 2 . 

4 x 2 /1 

8 x 2 /, 

!8 x */,. 18 x ^ 

20 x 3 /, 

5 x 2 /, . 

7 x *A, 

X "72, 

17x 3 / 2 . 

21 x 2 /, 

22 x %. 

23 x 3 / 2 . 


26 x 2 /i 

2 9 x 3 /, 


AS. Searle , del et lith , 

2 7. 



Trans EcySoaEduf 7 hind: lamellibr 



3 9 x %. 

45x 2 /i. 

51 x 5 A . 


40 x % . 

41x 3 / 2 . 

47 x 2 A 

52 x 5 /1 

43 x 2 /,. 

42 x % . 

48 x 3 / 2 . 

53 x 5 /, 

44 x a /i. 

48 x%. 

50 x %. 

54 x %. 

55 x 2 /, 

O. Searle lel.etMi. 

X 72. 

57 x % 

58 x%. 

58 x 3 /, 

( 361 ) 

XVI. —On a New Species of Dineuron and of Botryopteris from Pettycur, Fife. 
By R. Kidston, LL.D., F.R.S., F.R.S.E. (With One Plate.) 

(Read May 4, 1908. MS. received same date. Issued separately August 25, 1908.) 

Among the interesting specimens which have been yielded by the material of 
Calciferous Sandstone age (Culm), from Pettycur, near Burntisland, are a petiole of 
Dineuron and a very small species of Botryopteris. The former genus does not appear 
to have been previously discovered in Britain, and the present species is, as far as I am 
aware, only represented by a single example. The Botryopteris, in the form of 
fragments of petioles, is not infrequent, but its stems are of much more rare occurrence. 

I. Dineuron ellipticum, Kidston, n. sp. (Plate, figs. 1-3.) 

The specimen which forms the subject of the following description consists of a 
single transverse section of an almost circular petiole, whose greater width is about 
2*25 mm. (fig. 1). 

The petiole possesses an outer zone of stout cortex about 0'60 mm. wide (fig. 1, o.C), 
which is succeeded inwards by a narrow band of delicate inner cortex (fig. 1, in.C). 
This was separated by an endodermis (fig. 1, end.) from the thin-walled elements of 
the stele, which have almost entirely disappeared, and the mass of xylem now lies on 
one side of the space which they originally occupied (fig. 1). 

The xylem of Dineuron ellipticum consists of an elliptic mass whose greater 
diameter is about 070 mm. It is composed of large tracheae without any admixture of 
parenchyma ; towards its two extremities the tracheae suddenly become smaller where they 
meet the protoxylem elements (fig. 2, prx.). 

At the right side of the xylem mass, a short distance within its margin, is a circular 
opening surrounded by the protoxylem elements. According to Renault, this circular 
opening was originally filled with parenchyma,* which also occurs in a similar position 
in the stele of Zygopteris duplex, Will., sp.t \ 

At the left side of the stele a semicircular sinus is observable. This results from 
the separation of a portion of the xylem to form the outgoing pinna trace. There is, 
unfortunately, no clear evidence as to the mode of departure of the pinna trace, for the 
structure so interpreted by Renault in his Dineuron pteroides\ is more probably an 
unequal division of the petiole stele. One can only suggest, from the great similarity of 
the stele of Dineuron to that of Zygopteris duplex, that a band was cut off alternately from 

* In Dineuron pteroides, Renault, Bassin houil. et perm. d'Autun et d'Spinac, Flore foss., deux, part, p. 22, 1896. 

t Rachiopteris duplex, Will., Phil Trans., vol. clxiv., p. 687, Plates. 

| Several slides of Zygopteris duplex in my collection show this : Nos. 1315, 1314, 1313, etc. 

§ Renault, I.e., p. 23, fig. 19. 



each end of the xylem to supply the pinnae traces, which, as in Zygopteris duplex, 
possibly divided into two in their course through the cortex. 

The metaxylem is multiseriate as seen in transverse sections of the trachea?, while the 
protoxylem is apparently scalariform. 

The soft elements of the stele have all decayed, and are only represented by carbon- 
aceous fragments around its periphery. 

The remains of the endodermis are seen at fig. 1 , end. , and outside of this lies the 
inner cortex, composed of four or five rows of delicate parenchyma. This is succeeded 
by the outer cortex, formed of thick- walled prosenchyma, the component elements of 
which vary somewhat in the size of their lumen, but there is no regular arrangement of 
the larger and smaller elements. The peripheral portion of the cortex consists of smaller 
elements, but it is much destroyed and seems to have been of a more delicate structure 
than the zone lying immediately within it. 

If one compares the description of Dineuron ellipticum with Renault's description 
of his Dineuron pteroides, the wisdom of placing the Pettycur plant in the genus 
Dineuron may at first sight be questioned ; but if the figure of Dineuron pteroides be 
carefully examined, it will be seen that the tissue which occupies the central portion of 
the stele has much more the appearance of xylem than of parenchyma, as supposed by 
Renault. This circumstance, supported by the structure of the Pettycur plant, which 
agrees so completely with Dineuron in other respects, has led me to adopt this view, 
and a similar opinion seems to have been accepted by Mons. Paul Bertrand in his 
proposed classification of the Zygopteridese* 

II. Botryopteris antiqua, Kidston, n. sp. (Plate, figs. 4-12.) 

Three stems of Botryopteris antiqua are shown on Plate I. figs. 4, 6, 7. The one 
given at fig. 4 shows two attached petioles, and that at fig. 7 a single attached petiole, 
while the stem at fig. 6 illustrates the branching of the stele (s.', s") and also shows a 
petiole just freed from the stem. 

Characteristic of the family to which it belongs, Botryopteris antiqua has a very 
small stem stele, when considered in relation to the size of the petioles to which it gives 
rise. The stele of the specimen seen at fig. 4 is only # 40 mm. in diameter, while those 
of figs. 6 and 7 have a diameter of 0'50 mm. 

The stem is irregular in form, owing to the departure of the petioles and roots. The 
cortex is formed of thick-walled prosenchymatous cells of small diameter and without 
intercellular spaces, the larger elements being placed towards the outer surface of the 
stem, which bears numerous hairs formed of a single row of cells. 

The circular stem stele is formed of very small tracheae, without any admixture of 
parenchyma. Stem protoxylems are not distinguishable (fig. 5). In longitudinal 

* " Classification des Zygopteridees d'apres les caraeteres de leurs traces foliaires," Comptes rendus, 4th November 


section the tracheae are seen to possess scalariform thickenings without the occurrence of 
any porose markings. 

None of the stems yet met with show the departure of the leaf-trace from the stele, 
but in several of them the leaf- traces are seen in their passage through the cortex. 

At fig. 4, two petioles are being given off, apparently in spiral series ; at figs. 6 and 
7, one is seen on each stem, though in the former case it has probably become free. 

The petioles, which are larger than the stems from which they arise, are circular or 
slightly oval in form, the largest met with attaining a diameter of 2*20 mm., with a leaf- 
trace 075 mm. in its greatest width. The leaf- trace is thus a half larger than the 
diameter of the stele of the stem. 

The cortex of the petioles forms a broad zone of thick- walled prosenchymatous cells, 
the smaller of which lie at the periphery, while the larger elements are towards the 
centre of the zone, whence they decrease in size towards the endodermis, though even 
here they are larger than those of the peripheral area (fig. 6, pet.). 

The endodermis is clearly defined by its dark contents, and is seen at fig. 8, end. 

The whole of the soft elements of the leaf-trace have almost entirely disappeared, 
and when any fragments remain they are too imperfectly preserved to admit of a 
detailed description. The leaf-trace thus comes to occupy the centre of a clear space, or 
has fallen to one side (fig. 4, pet/, pet." ; fig. 6, pet.; fig. 9). 

In form the leaf-trace is oval, with one side slightly flattened, its more pointed or 
adaxial side being formed of protoxylem elements (fig. 8, prx.). In their distribution, 
however, the protoxylem elements are very irregular, and though their usual condition 
is to form a prominent band-like group at the apex of the trace, they occasionally have 
a greater or less lateral extension in the form of a narrow band down both sides of the 
xylem, or even extend some distance on the abaxial surface (fig. 8, prx.). At other 
times they are almost entirely limited to the adaxial margin, with a few isolated elements 
scattered along its lateral margins ; but in no case do the protoxylems form teeth as in 
the other known species of the genus. 

In longitudinal section the protoxylem elements are seen to be scalariform. 

The metaxylem of the leaf-trace is composed of large tracheae arranged without any 
definite order, but which become slightly smaller as they abut on the protoxylem 
(fig. 8). Their walls are porose. 

The petioles underwent bifurcation. At fig. 1 1 an early stage in the dichotomy of 
a petiole is seen. Here the trace has become transversely elongated, and an indenta- 
tion has appeared on one side. At fig. 12 the same petiole trace is seen divided into 
two equal arms. 

The pinnae appear to have been alternate, and their traces arise as small protuberances 
on the trace of the petiole (fig. 9, pin.). A further stage in their departure is seen at 
fig. 10, pin., where the pinna trace is free from that of the petiole, though still enclosed 
in a common cortex. The protoxylem of the pinna trace is adaxial to the trace of the 


The roots which rise direct from the xylem of the stem (fig. 6, r.) are very small, and 
contain a typical diarch strand composed of scalariform tracheae. 

The foliage and fructification of Botryopteris antiqua are unknown. 

Botryofiteris antique is a- typical member of the genus, though perhaps its smallest 
species, and is easily distinguished by its minute size and the protoxylem elements of 
the leaf-trace being evenly distributed and not forming prominent teeth as in the other 
known species. The tracheae of the stem are scalariform, not porose. 

The general character of growth of Botryopteris antiqua was that of a fern with a 
branching stem of slender dimensions which bore petioles of large size when compared 
with the stem from which they originated, and that must have supported itself by 
scrambling amongst the surrounding vegetation. 

It is an interesting point to notice the progressive development of the protoxylem 
in the form of prominent teeth. In Botryopteris antiqua, from the Calciferous Sand- 
stone series (Culm), protoxylem teeth are absent ; in Botryopteris hirsuta, Will., sp., 
from the Lamarkian series,* they are very distinct, and in some specimens even pro- 
minent ; while in Botryopteris forensis, Renault, from the Upper Coal Measures of Grand- 
Croix, they form long, narrow prolongations. There seems to be a tendency in the 
petiole trace to become more simple in form as traced back in geological time. 


[s.= stele; prx. = protoxylem ; in.C = inner cortex; o.C. = outer cortex; end. — endodermis ; pet. 
petiole ; r. = root ; h. = hairs ; pin. = pinna.] 

Figs. 1-3. Dineuron ellipticum, Kidston, n. sp. 

Fig. 1. Trans, section of petiole. x 30. Slide No. 57. 

Fig. 2. Trans, section of stele. x 70. 

Fig. 3. Trans, section of protoxylem group, x 160. 

Figs. 4-12. Botryopteris antiqua, Kidston, n. sp. 

Fig. 4. Trans, section of stem giving off two petioles. x 30. Slide No. 1084. 

Fig. 5. Trans, section of stele of last specimen. x 70. 

Fig. 6. Trans, section of stem showing division of stele, petiole, roots, and hairs. x 15. Slide No. 549c. 

Fig. 7. Trans, section of stem giving off a petiole x 15. Slide No. 782. 

Fig. 8. Trans, section of leaf-trace. x 80. Slide No. 508. 

Fig. 9. Trans, section of petiole giving off pinna. x 30. Slide No. 1347. 

Fig. 10. Trans, section of petiole giving off pinna. x 30. Slide No. 1086. 

Fig. 11. Trans, section of petiole showing early stage of dichotomy of leaf-trace, x 30. Slide No. 1348. 

Fig. 12. Trans, section of petiole showing leaf-trace dichotomously divided, x 30. Slide No. 1346. 
All the figured specimens are in the Author's collection. 

* Kidston, Quart. Journ. Geol. Soc., vol. lxi., p. 320, 1905. 

■Ins. Roy. Soc. Edin? 

Kidston : New Species of Dineuron and Botryopteris. 

Vol. XLVI. 

Fl g s - i-3— Dineuron ellipticum. Kidston, n.sp. 
Figs. 4-12.— Botryopteris antiqua. Kidston, n.sp. 

( 365 ) 

XVII. — The Craniology, Racial Affinities, and Descent of the Aborigines of Tasmania. 
By Principal Sir Wm. Turner, K.C.B., D.C.L., F.R.S. (With Three Plates.) 

(Read July 6, 1908. Issued separately October 16, 1908.) 


Introduction 365 

Description of Tasmanian skulls .... 368 

Comparison with Tasmanian skulls in other col- 
lections ........ 373 

Comparison of Tasmanians with other races . 381 


Racial affinities and descent of the Tasmanians . 385 

Sagittal contours ....... 394 

Bibliography 400 

Explanation of Plates and Figures in Text . . 403 


The Anatomical Museum of the University of Edinburgh contains a valuable collection 
of the skulls of the aborigines of Tasmania, which has not as yet been described. As 
the skulls of this now extinct people are limited in number in museums, and as the 
opportunity of collecting additional specimens no longer exists, I have thought that an 
account of their characters, a detailed statement of their measurements on lines similar 
to those pursued in my previous craniological memoirs, and a comparison of their con- 
formation with that of the Tasmanian skulls in other collections, as described by 
previous writers, would be of interest to anthropologists, and might assist in the pre- 
paration of a summary of their most constant features. Consideration of the affinities 
and possible descent of the Tasmanians may also be appropriately included in the Memoir. 

The collection began to be formed about the end of the first quarter of the last 
century. The first specimen in course of time — an adult male — was acquired by 
Professor Alexander Monro tertius (xxx. 1). He referred to it as a skull from Van 
Diemen's Land in his Elements of the Anatomy of the Human Body* in a chapter 
entitled, " On the distinctions in the skull of the male and female, and of the distinctions 
of the skulls of different nations." 

During the tenure of office of his successor, Professor John Goodsir, additional 
Tasmanian skulls were acquired for the Anatomical Museum. One of these, an adult 
male, is marked Van Diemen's Land, but with no other history (xxx. 5). Another, an 
aged edentulous male (xxx. 4), was presented by C. Gray, Esq. The skull-cap had 
been previously sawn off for the removal of the brain. The outer table of the parietal 
bones, at and near the middle of the sagittal suture, showed a large eroded patch, and 

* Second edition, p. 196, Edinburgh, 1831. Monro tertius died in 1859. I became acquainted with him in 1854, 
about which time he was having photographs made of the most interesting skulls in his collection. I possess a photo- 
graph of the skull, No. 52, referred to in the text, marked Van Diemen's Land in Monro's handwriting. Three 
measurements of the skull are given in Table ii., p. 204, of his Elements of Anatomy, but they are incorrectly 



on other parts of the parietal and frontal bones numerous shallow indentations were pre- 
sent which somewhat modified the contour and the general appearance of the cranium. 
They were possibly due to pathological conditions, or may have been produced by blows 
of the Tasmanian weapon named the waddy. # An adult male skull (xxx. 6) was 
presented as from Tasmania, to the late Professor Goodsir, shortly before his death in 
1867. The bones were discoloured, as if the skull had been buried, and the outer table 
at the vertex and on the right parietal was abraded and the diploe was partially exposed. 

During my incumbency of the Chair of Anatomy, other specimens were obtained. 
In 1870 an imperfect skull (xxx. 8), consisting of the frontal, both parietals, the supra- 
inial part of the occipital, the left temporal, and the right malar, was given by Mr J. 
Grant ; it was marked " extinct race V.D.L.," and from its appearance had probably been 
buried. Another imperfect specimen, marked " skull of an aborigine found at Bridge- 
water, presented by Mr Brent" (xxx. 9), consisted of the frontal and right parietal 
bones, obviously those of a young person. In October 1888 one of my pupils, Dr 
Lloyd H. Oldmeadow, presented to me the skull of an adult aboriginal male (xxx. 2) 
which he had brought from Hobart. It had been given to him by Dr E. M. Crowther 
of that town, and had been in the collection which had belonged to his father, Mr W. L. 
Crowther ; it was believed to be the skull of one of the last of the aborigines, and indeed 
possibly that of William Lanne, the last male to survive. In February 1889 an 
adult skull marked Tasmanian was given to me by Mr J. C. Robertson. It had 
previously been in the possession of Mr Seal, a member of one of the earliest 
families to settle in Tasmania, and was regarded by him as that of an aborigine ; it has 
female characters, and is marked xxx. 3 in Table I.t 

The collection in the Phrenological Museum of the Henderson Trustees, now lodged in 
the Anatomical Museum of the University, contains an adult male skull which is marked 
Van Diemen's Land. It is numbered 231 in the manuscript catalogue of that collection, 
compiled in 1858, though it had undoubtedly been in the collection some years before 
that date. Its number in the catalogue of the Anatomical Museum is xxx. 7. 

About the time when Monro tertius obtained the skull from Van Diemen's Land 
already referred to, Professor Robert Jameson had in his Museum of Natural History 
the skull of an aborigine marked Van Diemen's Land, which was also examined by Monro, 
who gave some measurements in Table ii., p. 204, in his chapter on the distinctive 
features of the skulls of different nations. The contents of Jameson's great museum 
were transferred to a department of the State in 1854, and they are now lodged in the 
Royal Scottish Museum, Edinburgh. The characters of the skull are embodied in 
the following description. 

* The waddy, about 2 feet long, was made of hard, heavy wood, sometimes knobbed at one end, which could be 
used as a club, or could be thrown with a rotary motion cither in battle or in the pursuit of prey. Barnard Davis, 
in his Thesaurus Craniorum and Supplement, refers to four Tasmanian skulls in his collection which showed marks of 
injury on the vault, the character of which is not specified, but may have been due to blows from the waddy. 

t The Tasmanian crania are Group xxx. in the Catalogue of the collection of Crania in the Anatomical Museum 
of the University. The specimens in each group have consecutive numbers. 



Table I. 

Tasmanian Crania in Edinburgh Museums. 
Group XXX. 

Collection number, - . 


Sex, .... 

Cubic capacity, . 

Glabello-occipital length, 

Basi-bregmatic height, 

Vertical Index, 

Minimum frontal diameter 

Stephanie diameter, 

Asterionic diameter, 

Greatest parieto-squamous 
breadth, . 

Cephalic Index, . 

Horizontal circumference, 

Frontal longitudinal arc, 

Parietal, ,, ,, 

Occipital, ,, ,, 

Total, ,, ,, 

Vertical transverse arc, 

Basi-transverse diameter, 

Vertical transverse circum 
ference, . 

Length of foramen magnum 

Basi-nasal length, 

Basi-alveolar length, . 

Gnathic Index, . 

Total longitudinal circum 
ference, . 

Interzygomatic breadth, 

Intermalar „ 

Nasio-mental length, 

Nasio-mental complete facial 

Nasio-alveolar length, . 

Maxillo -facial Index, . 

Nasal height, 

Nasal width, 

Nasal Index, 

Orbital width, 

Orbital height, 

Orbital Index, 

Palato-mnxillary length, 

Palato-maxiilary breadth, 

Palato-maxillary Index, 

Nasio-malar Index, 

Craniofacial Index, 
'Symphysial height, 
Coronoid „ 

Condoloid ,, 
Gonio-symphysial length 
Inter-gonial width, 
Breadth of ascending 

^ ramus, 




























69 4 
































































































59 6 









110 9 































71 1 




































































Description of Tasmanian Skulls. (Plates I. -III.) 

Measurements of the skulls are given in Table I., in which they are discriminated by 
the catalogue numbers, Group xxx. No. 1, e.s. With two exceptions they were adult 
males, though two were advanced in years, xxx. 3 was probably that of a woman, and 
the fragmentary calvaria xxx. 9 was that of a youth. The lower jaw was present in only 
two specimens, in one of which, the aged xxx. 4, it was edentulous. No definite 
statement can be made of the parts of the island in which the majority of the skulls 
were procured, but two (xxx. 2, 9) were obtained in or near Hobart Town. 

Norma verticalis. — The skulls were elongated, and dolichocephalic in the propor- 
tions of length and breadth. In three specimens the outline of the cranium was ovoid, 
but in the others the parietal eminences were so prominent, more especially in xxx. 2,7, 
that they broke the uniformity of the sides of the cranium, the outline of which approxi- 
mated to the pentagonal form. Behind the eminences the width of the cranium rapidly 
diminished into the occipital region. 

The frontal eminences were distinct. The frontal bone in the males showed in front 
of the bregma an area almost triangular in form, the broad base of which was at the 
coronal suture, whilst the apex approached a point between the frontal eminences ; its 
surface was convex from side to side and from base to apex. It was bounded 
laterally by a shallow, concave depression which extended backwards across the suture 
on to the parietal bone, as far as or somewhat beyond the parietal eminence. This 
depression was only feebly indicated in the single female skull in the collection. The 
temporal curved line was well marked in the male crania, and its anterior end formed 
the outer and lower boundary of the depression on the frontal bone. In two specimens 
the temporal line arched in the parietal region immediately above the eminences, and 
partially divided the depression on the vault into an upper and a lower area, the upper 
of which was the larger ; but in the other skulls it intersected the eminence at or near 
the greatest projection and formed the lower boundary of the depression. The width of 
the cranium in the frontal region, as compared with the parietal or parieto-squamous 
diameter, was relatively small. The mean Stephanie diameter was 103*2 mm. ; the 
mean parieto-squamous diameter was 133 mm. ; as the zygomatic arches were visible 
in the norma verticalis, the skulls were phaanozygous. 

The crania along the line of the sagittal suture were keeled, especially in its 
anterior half, though in some specimens in almost its whole length. In three skulls 
from 3 to 4 mm. behind the bregma the sagittal suture was depressed in a groove which 
was bounded on each side by a ridge which formed the upper boundary of the parietal 
depression ; the groove widened as it passed backwards to the lambdoid suture. In the 
other skulls the groove, with its lateral bounding ridges, was either absent, or so faintly 
marked as to be scarcely perceptible, and the upper boundary of the parietal depression 
was formed by the sagittal keel itself. The keel, conjoined with the steep, lateral slope 
of the parietal bones down to the eminences, gave a definite, roof-shaped character to the 



skulls, which, together with the shallow antero-posterior depressions in the fron to- parietal 
regions and the prominent parietal eminences, formed some of the distinctive features 
in the Tasmania